Merge pull request 'Let the harness declare what it runs on the host' (#141) from harness-host-exec into harness-guard-registration

Reviewed-on: #141
Reviewed-by: branderson <brad@branderson.io>

.dejima.md

@@ -123,3 +123,71 @@ the code was being worked on:
 Both are fixed by stating the precondition (`env -u`, a stubbed `ss`) rather
 than assuming it. When a test fails only in one environment, suspect this before
 suspecting the code — and when writing one, set what you depend on.
+
+## What lands on `main` here is a trust anchor, not just a release
+
+`docs/responsible-use.md` is this project's own policy, and two of its points
+bind while you are working *on* dejima rather than merely with it.
+
+**`.dejima/allowed-domains` is read from `origin/main` by
+dejima-ci-hardener**, which treats whatever it finds there as already
+reviewed - that is how a CI runner decides what a job may reach. Adding a
+domain to that file is not a convenience for one branch; it widens what every
+future CI job on that runner can talk to, from the moment it merges. Say what
+a domain is for in the pull request, and expect the question.
+
+**Security here is not a later pass.** This repository's product is an
+argument that a boundary holds. A change that quietly weakens one - a mount
+that widens, a check that moves after the thing it guards, a test that stops
+asserting what its name claims - is a defect in the product, not a rough edge
+to tidy up afterwards. If you notice one while doing something else, say so
+rather than leaving it for a pass that may not come.
+
+## Write the comment that prevents the bug, not the one that shows your work
+
+The prose here gets a compression pass periodically, and the same things get
+cut every time. Save that pass the work:
+
+- **Keep** the issue number, the failure mode, and why a check is shaped the
+  way it is. Someone has to be able to not reintroduce the bug.
+- **Cut how you found it.** Which CI run failed, how many attempts it took,
+  what you suspected first: that is your story, not the reader's.
+- **Cut the second and third consequence.** State the first. A reader who
+  needs the chain can follow it.
+- **Cut anything the code directly below already says.**
+- **Cut meta-commentary** - "the useful way to read this is", "what follows
+  is why". Say the thing instead.
+- **No numbers that drift.** Test counts, directory sizes, how many files a
+  sweep found: wrong within a month, and nothing checks them. This applies to
+  pull request descriptions too, which are read once and then stop being
+  true.
+- **One place.** An edge case belongs in `docs/`, not also in the README -
+  the README is what someone needs on first contact.
+
+This is not an argument for thinner documentation: the note on voice in the
+README prefers thorough over polished, deliberately. It is an argument
+against padding. A paragraph carrying one fact should be one sentence.
+
+The blocks that survive every pass are the ones where the argument *is* the
+content: the `O_NOFOLLOW` open in `dejima-project-lib.sh`, the id-derivation
+anchoring in `dejima-audit`'s `ack`, `--check-origin` in `dejima-ttyd`.
+Length is not the problem; length carrying no fact is.
+
+## Target `development`, not `main`
+
+Branches here are opened against `development`, which merges to `main`
+periodically. `main` stays releasable while work lands quickly; nothing else
+about the workflow changes.
+
+`gitea-pr` defaults its base to the repository's default branch, so say so:
+
+```sh
+gitea-pr create --base development --title ... --body-file -
+```
+
+Forgetting is the likely mistake, and it is recoverable - a pull request's
+base can be changed after the fact, in the web UI or through the API. Check
+what you opened rather than assuming.
+
+Rebase onto `origin/development` too, for the same reason: rebasing onto
+`main` puts anything already merged to `development` back in your diff.

ATTRIBUTION.md

@@ -0,0 +1,66 @@
+# Attribution
+
+The code in the dejima repositories is GPL-3.0-or-later; see `LICENSE`. The
+artwork is not code and is not covered by it.
+
+This file is the canonical statement for the dejima marks. Other repositories
+link here rather than restating the terms.
+
+## Authorship
+
+All logos, assets, and artwork are human-made. No AI tools were used in their
+creation. The artist is anonymous.
+
+The palette is based on [gruvbox](https://github.com/morhetz/gruvbox) by Pavel
+Pertsev, which is MIT-licensed. The colour values are gruvbox's; the marks
+drawn with them are not.
+
+## What is covered
+
+Both marks, in every form:
+
+- the logo: the filled tiles and the outline variants, in both light-mode and
+  dark-mode renderings, at any resolution, and the favicon;
+- the wordmark: the solid and outline renderings, again in both modes.
+
+The terms attach to the marks rather than to this list, so renderings not
+listed here are covered too.
+
+## Terms
+
+The dejima name, logo and wordmark are trademarks of Brad Anderson **as they
+identify this software**. A trademark attaches to a mark as used for
+particular goods and services, never to a word in the abstract, and the claim
+here goes no further than that.
+
+Dejima is a real place - the island in Nagasaki harbour this project is named
+after, and rebuilt there today. Nothing in this file claims the word itself,
+its historical or geographic sense, or any use unconnected with this
+software. Writing about the island, naming something after it, or using the
+name in any way that is not about this project is outside these terms and
+always was.
+
+The marks are not licensed under the GPL; GPL-3.0 §7(e) permits declining to
+grant trademark rights. All rights not granted below are reserved.
+
+**You may, without asking:**
+
+- reproduce them unmodified to refer to dejima;
+- redistribute them unmodified as part of an unmodified copy of a repository
+  that ships them.
+
+**You may not, without written permission:**
+
+- modify the artwork: recolour, redraw, crop, animate, or fold it into
+  another mark;
+- use the artwork in your own work, or as the mark of any other project,
+  product or organisation;
+- use the name for software in a way likely to be taken as referring to this
+  project - which is the only sense in which the name is claimed at all;
+- use any of them so as to suggest that dejima endorses, sponsors or produced
+  your work.
+
+A fork that remains a copy of dejima may keep the files. A fork that diverges
+into a project of its own must replace them with a mark of its own.
+
+Copyright in the image files is reserved.

CONTRIBUTING.md

@@ -14,15 +14,20 @@ different, **fork it and make the change in your fork**. Nothing here depends
 on an upstream maintainer accepting a patch. The fork is yours to run, and
 running it is the point.
 
+## Branches
+
+Work is opened against `development`, which merges to `main` periodically.
+`main` is what a reader should expect to be releasable; `development` is
+where it happens first.
+
 ## Expect it to move
 
 Significant change is coming in the near future. Config keys, interfaces, the
 layout of the docs, and the boundaries between the CLIs are all still
 settling. A fork taken today will diverge from upstream quickly.
 
-That is worth knowing before you build anything on top of one. If you do fork,
-pinning to a commit you have read is more useful here than tracking a branch
-that is being rewritten under you.
+If you do fork, pinning to a commit you have read is more useful here than
+tracking a branch that is being rewritten under you.
 
 ## If you are running an agent against this repository
 

README.md

@@ -1,7 +1,31 @@
-# dejima
+<!-- Fixed, not preference-aware, and that is the decision rather than an
+     oversight. `prefers-color-scheme` reports the browser's setting, never the
+     site's, and this is read on a forge that renders dark whichever way the
+     browser is set - so anything conditional shows dark ink on a dark page to
+     a reader whose browser prefers light, and does it silently.
+     The suffix names the mode a file is FOR, not the colour of its ink:
+     -darkmode is the pale wordmark that belongs on a dark page. A mirror on a
+     light-themed host swaps the filename below for dejima-wordmark-lightmode.png,
+     which is committed alongside for exactly that. One line, on the mirror,
+     where someone can see whether it worked.
+     Every &nbsp; here is spacing, and it has to be a character rather than CSS
+     or a presentation attribute. Gitea's sanitizer strips `style` from the tag
+     and from any wrapper, `hspace`/`vspace` survive sanitizing only to be
+     ignored by browsers that dropped HTML4 presentation attributes, and a bare
+     <br> ends a line nothing follows, which browsers collapse to no height.
+     &nbsp; survives as a real U+00A0: on its own line it fills a paragraph
+     that would otherwise be empty, and immediately before the tag it sits
+     inline and indents the image by one space. -->
+&nbsp;
 
-**Agents work best unsupervised. Bound the blast radius instead, and assume
-the agent is hostile.**
+&nbsp;<img src="assets/dejima-wordmark-darkmode.png" alt="dejima" height="72">
+
+---
+
+&nbsp;
+
+**Agents work best without you babysitting them. Bound the blast radius, let them
+work, and assume they are hostile.**
 
 An agent with permission prompts turned off gets a lot done. It can also be
 turned against you mid-session, by a prompt injection or a poisoned
@@ -13,13 +37,13 @@ Runs a coding agent (Claude Code by default, or whatever command you point
 it at) inside a firewalled Docker container with
 `--dangerously-skip-permissions` (no permission prompts, all tool use
 auto-accepted), against a project bind-mounted from the host. The agent gets
-to work without stopping to ask; the container is what makes that a
-reasonable trade rather than a reckless one.
+to work without stopping to ask; the container is what makes that trade a
+reasonable one.
 
-Around that core is most of what running an agent unattended actually takes:
+Around that core is most of what running an agent unattended takes:
 
 - **An egress allowlist.** DNS and IP, IPv4 and IPv6. Anything not on the
-  list does not connect, and the agent is told so rather than left to guess.
+  list does not connect, and the agent is told so.
 - **Credentials that stay on the host.** Your Anthropic OAuth credential
   never enters the container: a host-side proxy holds it and the container
   gets a placeholder. Your SSH key doesn't either; the container sees an
@@ -28,6 +52,10 @@ Around that core is most of what running an agent unattended actually takes:
   and dependencies it needs in `.dejima/`. Declaring is not granting: the
   approval is a separate decision you make on the host, by hand, and the
   approval store is never mounted where a session could reach it.
+- **One project's session cannot reach another's.** `~/.claude` is not
+  mounted whole. A session gets its own project's transcripts, memory and
+  edit history and nothing else, so it cannot read what you did elsewhere or
+  write into the memory a future session in another project will believe.
 - **A PR and issue workflow.** `gitea-pr` and `gitea-issue` let a session
   open pull requests and issues against Gitea or Forgejo and pull review
   feedback back in, on a token scoped for exactly that. Merging stays a
@@ -35,12 +63,16 @@ Around that core is most of what running an agent unattended actually takes:
 - **CI hardening, because push access is code execution.** A session opens a
   pull request, the workflow fires, and your runner executes that commit
   before you've read a line of it. The container bounds what a session reaches
-  directly, not what it can arrange to have run elsewhere. So this one is
-  work on the runner: [docs/ci.md](docs/ci.md) for what it takes, and
+  directly, not what it can arrange to have run elsewhere. Preventing this
+  requires hardening your CI runner: [docs/ci.md](docs/ci.md) for what it
+  takes, and
   [dejima-ci-hardener](https://gitbox.branderson.io/dejima/dejima-ci-hardener)
   for a tool that does it.
-- **Logging.** Every DNS lookup and rejected connection is shipped to Loki,
-  along with the session transcript if you opt in.
+- **Logging, off unless you turn it on.** Point `[logging] loki_url` at your
+  own Loki instance and every DNS lookup and rejected connection is shipped
+  there, with the session transcript too if you separately opt in. Nothing
+  leaves the machine otherwise, and there is no endpoint of ours anywhere in
+  this.
 - **Remote access, opt-in and off by default.** Attach to a live session
   from a terminal or a browser, with a push notification when it wants you.
 - **Guard rails inside the session.** Hooks refuse obviously destructive
@@ -52,17 +84,37 @@ Around that core is most of what running an agent unattended actually takes:
   and git identity carry over; uid, home directory and project path match on
   both sides, so a session's transcript is readable from the host afterward.
 - **An exit-time audit** of the host surfaces a session could have written
-  to, so a change to your git config or `~/.claude` gets surfaced instead of
-  just happening.
+  to, so a change to your git config or `~/.claude` gets surfaced.
 
-Originally inspired by
-[Anthropic's reference dev container](https://github.com/anthropics/claude-code/tree/main/.devcontainer),
-reworked to run via `docker compose` directly and grown well past it since.
+**Early development. Not ready to be depended on.** Nothing here has been
+through outside review, an external audit, or adversarial testing by anyone
+but the person who wrote it. Every security property this describes is a
+claim awaiting scrutiny, not one that has survived it. Do not put this
+anywhere that being wrong about it would matter.
+
+**One maintainer, no support.** Nothing about this is promised: not a reply,
+not a fix, not a release. [SECURITY.md](SECURITY.md) has the detail, and is
+where to report a vulnerability rather than in a public issue.
 
 This repo only holds the tooling; it's not where any project lives. Point
 it at a project on disk; that project is bind-mounted into the container at
 runtime.
 
+A note from the developer about the voice in this repository:
+
+> dejima is developed in dejima, and that means the words are written rapidly
+> by AI agents. That applies to this README, the docs, the commit messages,
+> the pull requests, and the code comments. I periodically take passes at
+> smoothing out the AI voice in the docs, but it's inefficient to do with
+> every pull request. I'm prioritizing thorough documentation over human
+> voice. If you feel something sounds painfully "AI", it is.
+>
+> The artwork (logo, wordmark, etc.) is a different story. It's made by a
+> human, and no AI tools were used to create it. See
+> [ATTRIBUTION.md](ATTRIBUTION.md).
+
+[![Developed in dejima](assets/dejima-badge.svg)](https://gitbox.branderson.io/dejima/dejima)
+
 ## The name
 
 [Dejima](https://en.wikipedia.org/wiki/Dejima) was a small artificial island
@@ -112,8 +164,8 @@ road with no barrier.
 
 ## Using this responsibly
 
-dejima bounds what a session can reach. It has nothing to say about whether
-what that session wrote should exist. That part is yours.
+dejima bounds what a session can reach. Whether what that session wrote
+should exist is your call.
 
 **Check whether the project wants it first.** A growing number of projects
 state that they do not accept AI-assisted contributions. Fork-and-review is
@@ -194,10 +246,17 @@ every key, its default, and what it is for.
 The forge integration, the push/PR workflow, and the agent harness aren't
 hardcoded into dejima: each is a directory under `forges/`, `workflows/` or
 `harnesses/`, selected independently by `[core] forge`, `[core] workflow` and
-`[core] harness`. Ships today: `forges/gitea`, `forges/github`, `forges/none`,
-`workflows/worktree-pr`, `workflows/branch-push`, `workflows/local-only`,
-`harnesses/claude-code`. The defaults are `forges/none` and
-`workflows/local-only`: a fresh install commits locally and pushes nowhere.
+`[core] harness`. `harness` has no default and must be set, so a later
+release cannot change which one you get.
+
+Ships today:
+
+- **Forges:** `none`, `gitea`, `github`
+- **Workflows:** `local-only`, `branch-push`, `worktree-pr`
+- **Harnesses:** `claude-code`
+
+The defaults are `forge = none` and `workflow = local-only`: a fresh install
+commits locally and pushes nowhere.
 
 See [docs/providers.md](docs/providers.md) for the manifest contract, key by
 key, and the shipped providers as worked examples.
@@ -229,8 +288,8 @@ to `worktree-pr` or `branch-push`: a pushing workflow is refused at launch
 unless its forge resolves to somewhere to push to, so the forge is not
 optional once you choose one of them.
 
-A pushing workflow also needs a git push credential, and which one you give it
-is a decision worth making deliberately rather than by default:
+A pushing workflow also needs a git push credential. Which one you give it
+matters:
 
 - **`[core] ssh_key`, a key of the session's own**, belonging to an account
   that is not you. This is the one to set up. It gives the sandbox an
@@ -280,7 +339,7 @@ dejima ~/repos/some-project
 
 By default the container's entry point is bare `claude` (always running with
 `--dangerously-skip-permissions`, regardless of how it's invoked; see
-`docker/claude-wrapper.sh`), so a fresh clone of this repo works with no
+`docker/harness-wrapper.sh`), so a fresh clone of this repo works with no
 editor setup at all.
 
 The loop, in the part every workflow shares: point it at a project, ask for a
@@ -306,8 +365,8 @@ merge, are what actually stop a merge from happening without you.
 What holds across all three is where they stop: none of them puts work on
 your default branch. `local-only` never leaves the machine, and the pushing
 workflows stop at a branch or at a pull request. The last step is yours in
-every case, which is a property of how the workflows are arranged rather than
-something enforced against a session that does not want to cooperate.
+every case. That comes from how the workflows are arranged, not from anything
+enforced against a session that does not want to cooperate.
 
 **→ [docs/configuration.md](docs/configuration.md#the-session-briefing)** -
 the guards in full, and why they fail open.
@@ -315,12 +374,22 @@ the guards in full, and why they fail open.
 ### Configuring the session
 
 Running a different editor or application than the default (`--cmd`,
-`[core] command`, `[mounts] nvim_config`), and mounting extra writable host
-directories a session should have access to (`[mounts] read_write`), are
-both config-file changes rather than anything you set up per launch.
+`[core] command`, `[mounts] nvim_config`) is a config-file change, as are the
+extra host directories a session should reach: `[mounts] readonly` and
+`[mounts] read_write`, mounted at their own path.
 
-**→ [docs/configuration.md](docs/configuration.md)** - both, including which
-host paths a writable mount is refused over and why.
+For a one-off, `--extra-readonly PATH` and `--extra-read-write PATH` (short
+forms `--ro` and `--rw`, both repeatable) add a path for a single launch, for
+things like reading another checkout while migrating code out of it. They
+append to those same lists, so a path a configured mount would be refused
+over is refused here too.
+
+```sh
+dejima --ro ~/other-project ~/myproject
+```
+
+**→ [docs/configuration.md](docs/configuration.md#one-off-mounts-for-a-single-launch)** -
+both forms, and which host paths a mount is refused over and why.
 
 ## Config parity with your host Claude Code setup
 
@@ -344,8 +413,7 @@ the project mount, for passing files in and out of a session without
 touching the actual checkout. It persists across sessions since it's a real
 directory, not a container volume - but treat it as a handoff point, not
 general-purpose storage. **Clear it out between sessions.** Anything left
-there is readable by the next session, and by every session after that, which
-turns a file you forgot into something you handed over.
+there is readable by the next session, and by every session after that.
 
 `~/shared`, the project mount, and anything in `[mounts] read_write` are the
 only writable host paths in the container. Everything else a session writes
@@ -456,6 +524,14 @@ sources, opt-in conversation logging, and viewing both in Grafana.
 
 ## GitHub push: blocked by default, optional opt-in
 
+A session that is not being used to push to a forge should not be able to
+push to it. Your configured forge is the one you chose; GitHub is where
+everything else lives - every dependency a session reads, every repository it
+clones to understand an error. The case this prevents is an eager agent, not
+a compromised one: an unsolicited pull request on a stranger's project, in
+your name, added to a queue its maintainers are already drowning in. See
+[docs/responsible-use.md](docs/responsible-use.md) for what to do instead.
+
 The sandbox can always read from GitHub - clone, fetch, `gh` API calls. By
 default it can never push there: GitHub's IP ranges are only allowlisted on
 port 443, so port 22 is physically unreachable, not merely discouraged by a
@@ -519,11 +595,10 @@ you what changed.
 
 It is **advisory, not a block**. It fires after the session has ended, so a
 finding is something to read and clean up before your next `git` or `claude`
-in that repo, not something that was prevented. And its diff is time-windowed
-rather than session-attributed: it reports that a path changed while the
-session was running, not that the session changed it.
+in that repo. Its diff is also time-windowed: it reports that a path changed
+while the session was running, not that the session changed it.
 
-`dejima-audit ack` walks the findings and shows you each one before asking.
+`dejima-audit show` indexes what is waiting; `dejima-audit ack` walks it.
 Approving records the exact content, so any later change re-fires; ignoring
 records the surface, so that path is fine whatever it later contains; and
 skipping decides nothing, which means you are asked again next session.
@@ -551,7 +626,7 @@ credentials never enter the container, that a push to anywhere but the
 allowed remote is refused. Hundreds of tests across suites covering
 static analysis, host-side behaviour, and containers started for real.
 
-They are the reason claims here are stated flatly rather than hedged.
+They are why the claims here are stated flatly.
 
 → **[docs/development.md](docs/development.md)**: the suites, what each
 covers, and how to run them.
@@ -567,6 +642,8 @@ covers, and how to run them.
 | [ci.md](docs/ci.md) | What push access grants on your CI host, and how to bound it |
 | [remote-access.md](docs/remote-access.md) | Picking and verifying a bind address, and the SSH and web front doors |
 | [development.md](docs/development.md) | The file map, the test suites, running them |
+| [SECURITY.md](SECURITY.md) | What dejima claims, where to report a vulnerability, and what is promised in return |
+| [ATTRIBUTION.md](ATTRIBUTION.md) | Who holds the dejima marks and what may be done with them |
 
 ## License
 
@@ -574,3 +651,11 @@ This project is licensed under the GNU General Public License v3.0 or later
 (GPL-3.0-or-later); see [LICENSE](LICENSE) for the full text.
 
 Copyright (C) 2026 Brad Anderson
+
+The dejima wordmark and logo are not code and are not covered by that licence.
+They are trademarks; who holds them and what may be done with them is stated
+once for the organisation, in [ATTRIBUTION.md](ATTRIBUTION.md).
+
+The egress firewall's original shape came from
+[Anthropic's reference dev container](https://github.com/anthropics/claude-code/tree/main/.devcontainer).
+`docker/init-firewall.sh` records what it kept and what it did not.

SECURITY.md

@@ -0,0 +1,77 @@
+# Security policy
+
+This is the security policy for every repository in the dejima organisation.
+Individual repositories state what is specific to them and link here for the
+rest.
+
+## What dejima claims
+
+It bounds what an agent session can reach. It does not make an agent safe to
+run.
+
+A session runs in a container with no permission prompts, one bind-mounted
+project directory, a deny-by-default egress allowlist covering DNS as well as
+IP, and no copy of the credentials it authenticates with. That allowlist
+bounds where a prompt injection or a compromised dependency can send code and
+where it can fetch its next instruction. It does not prevent either one.
+
+Nothing here holds against a container escape or a kernel bug. The in-session
+hooks are fast-fail guards, not a boundary. The exit-time audit runs after the
+session has ended and attributes nothing to it. And none of it speaks to
+whether the code a session wrote is correct, secure, or fit to merge.
+
+[docs/security.md](docs/security.md) states the threat model and covers each
+of these in detail.
+
+## Reporting a vulnerability
+
+Email **dejima@branderson.io**. Please do not open a public issue for a
+vulnerability; the tracker is public, and a report there is a disclosure.
+
+Useful things to include:
+
+1. what an attacker gets;
+2. the steps to reproduce it;
+3. the repository and the commit you tested.
+
+## What you can expect in return
+
+Nothing is promised: no reply, no acknowledgement, no triage, no fix, and no
+timeline for any of them. Reports are read when there is time to read them.
+
+Nothing is asked in return either. There is no embargo period and no
+coordinated disclosure window. Disclose on whatever schedule you think is
+right.
+
+## This is not proven software
+
+These projects are in early development. Nothing in them has been through
+outside review, an external audit, or adversarial testing by anyone other
+than the people who wrote them. Where a design argument is stated it is one
+person's reasoning; where a test asserts something it asserts what that
+person thought to check.
+
+Treat every security property described in any of these repositories as a
+claim awaiting scrutiny rather than one that has survived it, and do not
+deploy them where being wrong would matter.
+
+## There is no support
+
+These are personal projects with one maintainer.
+
+There is no paid version, no support contract, no service level, no sponsorship
+attached to them, and no commercial relationship with anyone who uses them.
+Questions may go unanswered. Issues may go unread. Nothing here is developed or
+supplied in the course of a commercial activity.
+
+Whoever deploys this is responsible for the result.
+
+## Supported versions
+
+The tip of `main`, and nothing else. There are no maintenance branches, no
+backports and no security releases for older tags or commits.
+
+## No warranty
+
+The software is provided without warranty of any kind. See sections 15 and 16
+of each repository's `LICENSE`.

assets/README.md

@@ -0,0 +1,19 @@
+# assets
+
+The dejima wordmark, in the two versions the README needs: `-darkmode` for a
+page that renders dark, `-lightmode` for one that renders light. The suffix
+names the mode a file is for, not the colour of its ink, so `-darkmode` is the
+pale one.
+
+The README points at `-darkmode` unconditionally. A mirror on a light-themed
+host changes that one line; the comment above the tag says why it is not
+conditional.
+
+**These are trademarks and are not covered by this repository's GPL licence.**
+Who holds them, and what may be done with them, is stated once for the whole
+organisation, in this repository:
+
+[ATTRIBUTION.md](../ATTRIBUTION.md)
+
+Other repositories in the organisation link to that file by absolute URL, so
+it does not move or get renamed without their links breaking with it.

assets/dejima-badge.svg

@@ -0,0 +1,21 @@
+<svg xmlns="http://www.w3.org/2000/svg" width="122" height="20" role="img"
+     aria-label="Developed in dejima">
+  <title>Developed in dejima</title>
+  <linearGradient id="g" x2="0" y2="100%">
+    <stop offset="0" stop-color="#bbb" stop-opacity=".1"/>
+    <stop offset="1" stop-opacity=".1"/>
+  </linearGradient>
+  <clipPath id="r"><rect width="122" height="20" rx="3" fill="#fff"/></clipPath>
+  <g clip-path="url(#r)">
+    <rect width="79" height="20" fill="#282828"/>
+    <rect x="79" width="43" height="20" fill="#8ec07c"/>
+    <rect width="122" height="20" fill="url(#g)"/>
+  </g>
+  <g font-family="Verdana,DejaVu Sans,Geneva,sans-serif" font-size="11"
+     text-anchor="middle">
+    <text x="39.5" y="15" fill="#010101" fill-opacity=".3">Developed in</text>
+    <text x="39.5" y="14" fill="#ebdbb2">Developed in</text>
+    <text x="100.5" y="15" fill="#010101" fill-opacity=".3">dejima</text>
+    <text x="100.5" y="14" fill="#282828">dejima</text>
+  </g>
+</svg>

bin/dejima

@@ -23,6 +23,10 @@ source "$SCRIPT_DIR/dejima-project-lib.sh"
 source "$SCRIPT_DIR/dejima-provider-lib.sh"
 # shellcheck source=dejima-color-lib.sh
 source "$SCRIPT_DIR/dejima-color-lib.sh"
+# shellcheck source=dejima-slug-lib.sh
+source "$SCRIPT_DIR/dejima-slug-lib.sh"
+# shellcheck source=dejima-harness-lib.sh
+source "$SCRIPT_DIR/dejima-harness-lib.sh"
 # fd 2: everything this script says for itself goes to stderr, so the
 # session's own stdout stays exactly what the container wrote.
 color_init 2
@@ -59,6 +63,65 @@ _need_value() {
     }
 }
 
+# Per-launch additions to [mounts] readonly and [mounts] read_write, for the
+# one-off cases: reading another checkout while migrating code out of it,
+# writing to a scratch directory for one session. Repeatable, one path each.
+#
+# These append to the config's own lists rather than carrying their own mount
+# code, which is the whole design. Everything that already applies to a
+# configured mount then applies to a flag automatically and with nothing
+# duplicated: the existence check, the refusal when a path names the approval
+# store or the audit stores, the refusal when the same path appears in both
+# lists, and the read-write paths the session briefing and the audit report
+# header name. A flag is a way to say where, not a way around any of that.
+EXTRA_RO_FLAGS=()
+EXTRA_RW_FLAGS=()
+FLAG_MOUNTS=""
+_add_mount_flag() {
+    local flag="$1" value="$2"
+    # Both lists are colon-separated, so a path containing one cannot be
+    # represented in them at all - it would silently split into two mounts
+    # that each fail the existence check. Say so instead.
+    case "$value" in
+        *:*) echo "$flag path must not contain a colon: $value" >&2; exit 1 ;;
+    esac
+    # Same leading-~ expansion the config's list:path type does, for the
+    # quoted case the shell leaves alone.
+    value="${value/#\~/$HOME}"
+    # Relative paths are worth accepting here and meaningless in a config
+    # file, which has no working directory to be relative to. Made absolute
+    # by prefixing rather than by readlink -f: resolving would mount the
+    # symlink's target under a path the operator never typed.
+    case "$value" in /*) ;; *) value="$PWD/$value" ;; esac
+    case "$flag" in
+        --extra-readonly)   EXTRA_RO_FLAGS+=("$value") ;;
+        --extra-read-write) EXTRA_RW_FLAGS+=("$value") ;;
+    esac
+    # Recorded here, at parse time, and not where the lists are assembled:
+    # that happens inside a command substitution, whose subshell would throw
+    # the assignment away and leave every flag reported as a config entry.
+    FLAG_MOUNTS="${FLAG_MOUNTS:+$FLAG_MOUNTS:}$value"
+}
+_append_mounts() {
+    local list="$1"; shift
+    local p
+    for p in "$@"; do
+        list="${list:+$list:}$p"
+    done
+    printf '%s' "$list"
+}
+# Which of the two the operator actually wrote, for the refusals below.
+# Without this a mistyped --ro is reported as a "[mounts] readonly entry",
+# sending someone to a config file that does not mention the path at all.
+
+_mount_label() {
+    local path="$1" from_config="$2" from_flag="$3"
+    case ":$FLAG_MOUNTS:" in
+        *":$path:"*) printf '%s' "$from_flag" ;;
+        *)           printf '%s' "$from_config" ;;
+    esac
+}
+
 PRINT_CONFIG=0
 PRINT_BRIEFING=0
 ALLOW_GITHUB_PUSH_FLAG=0
@@ -72,6 +135,12 @@ while [ $# -gt 0 ]; do
         --cmd)      _need_value --cmd "${2:-}";      _set_override core.command "$2";  shift 2 ;;
         --forge)    _need_value --forge "${2:-}";    _set_override core.forge "$2";    shift 2 ;;
         --workflow) _need_value --workflow "${2:-}"; _set_override core.workflow "$2"; shift 2 ;;
+        --extra-readonly|--ro)
+            _need_value --extra-readonly "${2:-}"
+            _add_mount_flag --extra-readonly "$2"; shift 2 ;;
+        --extra-read-write|--rw)
+            _need_value --extra-read-write "${2:-}"
+            _add_mount_flag --extra-read-write "$2"; shift 2 ;;
         --*) echo "unknown option: $1" >&2; exit 1 ;;
         *) break ;;
     esac
@@ -162,28 +231,18 @@ if ! config_exists; then
     exit 1
 fi
 
-# A selected forge's env_export can name a variable core also fills from its
-# own config. Today that is GH_TOKEN, which core sets from the read-only
-# credential at [github] token_file - one every session gets regardless of
-# which forge is selected - and which a forge exporting the same name takes over
-# for the launch, since a single variable can only carry one token. Said out
-# loud rather than left as two configured credentials with one silently
-# ignored: an operator who set both is entitled to know which is in effect. A
-# warning, not a refusal - the forge's value winning is exactly what
-# selecting that forge asked for.
+# A forge's env_export can name a variable core also fills - today GH_TOKEN,
+# which core sets from [github] token_file for every session regardless of
+# forge. One variable carries one token, so the forge's wins. Warned rather
+# than left silent: an operator who configured both is entitled to know which
+# is in effect. Not a refusal, since the forge winning is what selecting it
+# asked for.
 #
-# Driven off the manifest's env_export rather than off the forge's name: the
-# collision is a property of the variable a forge exports, identical for any
-# forge that exports it, and core has no business knowing which forge that is
-# today. Checked here, ahead of --print-config's own exit below (like
-# _warn_legacy_env above it) rather than down by the GH_TOKEN block further
-# down that --print-config never reaches, so --print-config reports it too.
-#
-# Both files, not just core's: a forge key resolving to nothing is the
-# "wholly unconfigured" case forge_env_exports deliberately emits no pair for
-# (see its own comment), leaving core's own value in effect - so claiming the
-# forge's key won would be false in exactly that state. The warning only
-# means something once both are actually set.
+# Driven off env_export rather than the forge's name - the collision belongs
+# to the variable, not to any particular forge. Checked here, ahead of
+# --print-config's exit, so that reports it too. Both values must actually be
+# set: a forge key resolving to nothing leaves core's in effect, and claiming
+# the forge won would be false.
 _warn_forge_env_collisions() {
     local forge="$1" pair name ref core core_name core_section core_leaf
     local core_file forge_file
@@ -239,8 +298,42 @@ fi
 DEJIMA_FORGE_NAME="$(config_get core forge)"
 DEJIMA_WORKFLOW_NAME="$(config_get core workflow)"
 DEJIMA_HARNESS_NAME="$(config_get core harness)"
+# No default: claude-code was one, and a release changing it would move an
+# operator who never chose to a different harness on upgrade.
+if [ -z "$DEJIMA_HARNESS_NAME" ]; then
+    echo "[core] harness is not set in $(config_path)." >&2
+    echo "There is no default, so a later release cannot change your choice." >&2
+    echo "" >&2
+    echo "  [core]" >&2
+    echo "  harness = claude-code" >&2
+    echo "" >&2
+    echo "Available: $(provider_list harness | tr '\n' ' ')" >&2
+    exit 1
+fi
 provider_validate_selection "$DEJIMA_FORGE_NAME" "$DEJIMA_WORKFLOW_NAME" "$DEJIMA_HARNESS_NAME" || exit 1
 
+# The selected harness's plugin: what dejima does for this harness that no
+# manifest value could state. Loaded once the selection is validated, and
+# before anything that calls a hook. A manifest without a plugin is refused
+# rather than treated as a lighter kind of support - see "Adding a new
+# harness" in docs/development.md.
+harness_plugin_load "$DEJIMA_HARNESS_NAME"
+
+# Whether GitHub is a push destination this session, which the briefing tells
+# the agent about. Two ways in: the per-launch flag, or a forge whose hosts
+# core allowlists (github) under a workflow that pushes. Derived from the
+# manifests directly rather than from DEJIMA_WORKFLOW_PUSH, which is resolved
+# further down - this has to be settled before --print-briefing exits, or that
+# would render a briefing a real launch does not.
+DEJIMA_GITHUB_PUSH=off
+if [ "$ALLOW_GITHUB_PUSH_FLAG" = 1 ]; then
+    DEJIMA_GITHUB_PUSH=on
+elif [ "$(manifest_get forge "$DEJIMA_FORGE_NAME" hosts_allowlisted_by_core)" = "on" ] \
+     && [ "$(manifest_get workflow "$DEJIMA_WORKFLOW_NAME" push)" = "on" ]; then
+    DEJIMA_GITHUB_PUSH=on
+fi
+export DEJIMA_GITHUB_PUSH
+
 # Resolved here, ahead of --print-briefing below, because briefing_compose
 # needs DEJIMA_RW_MOUNTS (the {{writable_mounts}} placeholder) to render the
 # same sentence a real launch renders - see the read_write mounts block this
@@ -266,6 +359,8 @@ WORKSPACE_DIR="$(cd "$WORKSPACE_DIR" && pwd)"
 AUDIT_DIR="$(config_get audit dir)"
 BASELINE_DIR="$(config_get audit baselines_dir)"
 SANDBOX_EXTRA_MOUNTS="$(config_get mounts readonly)"
+SANDBOX_EXTRA_MOUNTS="$(_append_mounts "$SANDBOX_EXTRA_MOUNTS" \
+    ${EXTRA_RO_FLAGS[@]+"${EXTRA_RO_FLAGS[@]}"})"
 
 # Read-write host mounts. Distinct from [core] share_dir, which is mounted at
 # ~/shared - a well-known handoff point the agent is told about by name -
@@ -286,7 +381,41 @@ _paths_overlap() {
     case "$b/" in "$a"/*) return 0 ;; esac
     return 1
 }
+
+# The stores that decide what a session may do next, and the record of what
+# it did. Read-only is not enough to make mounting them harmless: the
+# approval store discloses every project's approved domains and
+# dependencies, and the audit baselines say which findings have already been
+# waved through. Refusing beats mounting and hoping the mode holds (#15).
+GUARDED_STORES=()
+_guarded_stores() {
+    [ ${#GUARDED_STORES[@]} -eq 0 ] || return 0
+    GUARDED_STORES=("$(config_root)" "$AUDIT_DIR" "$BASELINE_DIR")
+}
+
+# Refuse $2 if it overlaps any of the paths in $3.., in either direction: an
+# ancestor of a store exposes it just as surely as the store itself.
+_refuse_guarded_mount() {
+    local kind="$1" path="$2"; shift 2
+    local real guarded guarded_real
+    # readlink -m rather than -f, same reason as the read_write guard below:
+    # -f gives nothing for a path that does not exist yet, and an empty
+    # result compares equal to nothing and would sail through.
+    real="$(readlink -m "$path")"
+    for guarded in "$@"; do
+        [ -n "$guarded" ] || continue
+        guarded_real="$(readlink -m "$guarded")"
+        _paths_overlap "$real" "$guarded_real" || continue
+        echo "Refusing $kind: $path" >&2
+        echo "It overlaps $guarded_real, which dejima needs out of a session's reach." >&2
+        echo "Even read-only, that hands the session every project's approvals or the" >&2
+        echo "record of what earlier sessions were caught doing." >&2
+        exit 1
+    done
+}
 SANDBOX_RW_MOUNTS="$(config_get mounts read_write)"
+SANDBOX_RW_MOUNTS="$(_append_mounts "$SANDBOX_RW_MOUNTS" \
+    ${EXTRA_RW_FLAGS[@]+"${EXTRA_RW_FLAGS[@]}"})"
 RW_MOUNT_PATHS=()
 DEJIMA_RW_MOUNTS=""
 if [ -n "$SANDBOX_RW_MOUNTS" ]; then
@@ -294,7 +423,7 @@ if [ -n "$SANDBOX_RW_MOUNTS" ]; then
     for path in "${rw_paths[@]}"; do
         [ -n "$path" ] || continue
         if [ ! -d "$path" ]; then
-            echo "[mounts] read_write entry is not an existing directory: $path" >&2
+            echo "$(_mount_label "$path" "[mounts] read_write entry" "--extra-read-write path") is not an existing directory: $path" >&2
             exit 1
         fi
         # readlink -m, not -f: -f yields nothing unless at most the final
@@ -311,7 +440,7 @@ if [ -n "$SANDBOX_RW_MOUNTS" ]; then
             "$HOME/.claude" "$HOME/.local/bin" "$REPO_ROOT" "$WORKSPACE_DIR"; do
             guarded_real="$(readlink -m "$guarded")"
             _paths_overlap "$rw_real" "$guarded_real" || continue
-            echo "Refusing [mounts] read_write entry: $path" >&2
+            echo "Refusing $(_mount_label "$path" "[mounts] read_write entry" "--extra-read-write path"): $path" >&2
             echo "It overlaps $guarded_real, which dejima needs out of a session's reach." >&2
             echo "A writable mount there would let a session approve its own domains and" >&2
             echo "dependencies, delete the audit findings it is named in, or plant code" >&2
@@ -508,12 +637,20 @@ fi
 # A valid forwarded socket with zero identities loaded forwards fine (the
 # container sees a live, working agent) but is useless for git-over-ssh -
 # warn now rather than have that surface confusingly deep into a session as
-# a failed push. Non-fatal: a task might not need git-over-ssh at all, and
-# SANDBOX_SSH_KEY (if set) provides its own identity independently of this.
-if [ "$have_forwarded_agent" = 1 ] && ! ssh-add -l >/dev/null 2>&1; then
+# a failed push. Non-fatal: a task might not need git-over-ssh at all.
+#
+# Silent when [core] ssh_key is set, because then the premise is false: the
+# key is loaded into the container's own agent and pushes work without the
+# forwarded one carrying anything. It is also the better arrangement of the
+# two - a dedicated identity that works unattended, rather than the
+# operator's own - so warning about the fallback being empty, while the
+# recommended path is configured and working, tells someone who took the
+# advice that they got it wrong.
+if [ "$have_forwarded_agent" = 1 ] && [ -z "$SANDBOX_SSH_KEY" ] \
+   && ! ssh-add -l >/dev/null 2>&1; then
     echo "Warning: ssh-agent has no identities loaded (ssh-add -l shows none)." >&2
     echo "git-over-ssh (e.g. pushing to the configured forge) will fail until you ssh-add a key," >&2
-    echo "unless [core] ssh_key is also set." >&2
+    echo "or set [core] ssh_key in $(config_path) to a dedicated key - see docs/configuration.md." >&2
 fi
 
 # Deliberately not `gh auth token` / gh's own hosts.yml: a normal `gh
@@ -570,47 +707,29 @@ if ! settings_error="$(jq . "$HOME/.claude/settings.json" 2>&1 >/dev/null)"; the
     exit 1
 fi
 
-# Claude Code's own internal per-command sandboxing (~/.claude/settings.json's
-# "sandbox" block) is redundant here - it can't even function inside this
-# container (bwrap can't create a namespace, and getting it to would need
-# privilege escalation approaching --privileged, undoing this container's own
-# hardening), and this container is already a real, working sandbox on its
-# own. A session can't be trusted to turn it off itself, so the override is
-# generated here, on the host, and reaches Claude Code as an extra settings
-# layer: docker-compose.yml mounts this file read-only at a container-local
-# path and claude-wrapper.sh passes it as --settings, which loads above
-# ~/.claude/settings.json in the precedence order. Everything else in the
-# operator's real settings (hooks, permissions, plugins) still applies from
-# there, so this file holds only what the sandbox overrides.
+# Claude Code's own per-command sandboxing is redundant here and cannot work
+# anyway - bwrap can't create a namespace without privileges that would undo
+# this container's hardening - and a session can't be trusted to turn it off
+# itself. So the override is generated host-side and reaches Claude Code as
+# an extra --settings layer above ~/.claude/settings.json, which holds only
+# what the sandbox overrides; the operator's hooks, permissions and plugins
+# still apply from there.
 #
-# The earlier arrangement - a full modified copy of the operator's settings,
-# mounted over ~/.claude/settings.json - detached itself whenever the host
-# rewrote the real file, because that mount point sits inside the read-write
-# ~/.claude mount and a rename unlinks the dentry the mount is pinned to. The
-# session then silently fell back to the operator's settings, with sandboxing
-# back on and the notify hooks below gone. Mounting outside ~/.claude removes
-# that: nothing on the host can rename over a container-local path. See #47.
+# It is mounted at a container-local path, not over ~/.claude/settings.json.
+# That earlier arrangement detached whenever the host rewrote the real file
+# (a rename unlinks the dentry the mount is pinned to), silently restoring
+# the operator's settings with sandboxing back on. See #47.
 SANDBOX_SETTINGS_FILE="$(mktemp "${TMPDIR:-/tmp}/dejima-settings.XXXXXX.json")"
 jq -n '{sandbox: {enabled: false}}' > "$SANDBOX_SETTINGS_FILE"
 
-# ~/.claude.json mixes live app state (project registry, usage stats,
-# caches) with config (mcpServers) that can't be split at the mount level -
-# unlike settings.json above, it can't just be made read-only either:
-# Claude Code does an atomic write-temp-then-rename against this exact file
-# during normal startup (before login even completes), and a read-only
-# file makes that hang instead of failing cleanly (confirmed against a
-# real read-only copy of this file - see
-# https://github.com/anthropics/claude-code/issues/29162). Generate a
-# throwaway copy instead, seeded from the real file, and mount *that*
-# read-write in its place - same "generate a sandbox-specific file on the
-# host" pattern as SANDBOX_SETTINGS_FILE above, just solving a different
-# problem: nothing a compromised session writes here (mcpServers or
-# anything else) ever reaches the real file, since the copy is discarded,
-# not merged back, when this session ends. This also means concurrent
-# sandbox sessions never contend over the same file - each gets its own
-# copy. See docs/configuration.md "Config parity with your host Claude Code
-# setup" for exactly what that means
-# doesn't carry over from a sandboxed session.
+# ~/.claude.json mixes live app state with config (mcpServers) that cannot be
+# split at the mount level, and read-only is not an option: Claude Code does
+# an atomic write-and-rename against it during startup, and a read-only file
+# makes that HANG rather than fail cleanly (anthropics/claude-code#29162).
+# So it gets a throwaway copy, seeded from the real file and discarded rather
+# than merged back - nothing a session writes here reaches the host, and
+# concurrent sessions never contend. See docs/configuration.md, "Config
+# parity with your host Claude Code setup".
 SANDBOX_CLAUDE_JSON_FILE="$(mktemp "${TMPDIR:-/tmp}/dejima-claude-json.XXXXXX.json")"
 if [ -f "$HOME/.claude.json" ]; then
     cp "$HOME/.claude.json" "$SANDBOX_CLAUDE_JSON_FILE"
@@ -637,11 +756,32 @@ SANDBOX_AUDIT_STATE="$(mktemp "${TMPDIR:-/tmp}/dejima-audit.XXXXXX")"
 # and SANDBOX_DERIVED_CONTEXT is one of the `rm -rf` arguments.
 DEJIMA_NOTIFY_DIR=""
 DEJIMA_SESSION_FILE=""
-SANDBOX_REMOTE_SETTINGS_TMP=""
 SANDBOX_NVIM_LOCKFILE=""
 SANDBOX_DERIVED_DOCKERFILE=""
 SANDBOX_DERIVED_CONTEXT=""
 
+# Names each host-side task in cleanup() as it starts. Between the container
+# exiting and the operator's shell prompt returning, dejima audits every
+# surface the session could have written to and reads back the transcripts it
+# left; on a real project that is seconds of nothing on screen, and a
+# terminal that prints nothing reads as one that has hung. The remedy an
+# operator reaches for then is Ctrl-C, which is precisely when the audit
+# must not be interrupted.
+#
+# Only once a snapshot exists, which is the same condition `dejima-audit
+# compare` uses to decide it has anything to do: an exit before the session
+# ever started is instant, and its last line is the error that caused it -
+# announcing work that is not happening would push that off the screen.
+#
+# stderr, like everything else this trap prints, and host-side only: none of
+# it is ever routed back into a session's context.
+exit_step() {
+    # `return 0` rather than letting the test be the last command: a nonzero
+    # status from anything in an EXIT trap replaces the script's own.
+    [ -s "${SANDBOX_AUDIT_STATE:-}" ] && printf 'dejima: %s...\n' "$1" >&2
+    return 0
+}
+
 cleanup() {
     # Every rm below is guarded with `|| true`: under set -e, a failing
     # command in an EXIT trap replaces the shell's own exit status, so none
@@ -673,7 +813,7 @@ cleanup() {
     fi
     rm -f "$SANDBOX_SETTINGS_FILE" "$SANDBOX_CLAUDE_JSON_FILE" "${SANDBOX_NVIM_LOCKFILE:-}" "${SANDBOX_FAKE_CREDENTIALS_FILE:-}" "${SANDBOX_DERIVED_DOCKERFILE:-}" "${SANDBOX_BRIEFING_FILE:-}" || true
     rm -rf "${SANDBOX_GITCONFIG_DIR:-}" "${SANDBOX_ANTHROPIC_PROXY_DIR:-}" "${SANDBOX_DERIVED_CONTEXT:-}" || true
-    rm -f "${DEJIMA_SESSION_FILE:-}" "${SANDBOX_REMOTE_SETTINGS_TMP:-}" || true
+    rm -f "${DEJIMA_SESSION_FILE:-}" || true
     # Unlike the silent `|| true` guards above, this removal (and, before
     # the audit below was moved in, the last command of an EXIT trap) warns
     # out loud when it fails: the marker directory is bind-mounted
@@ -685,6 +825,15 @@ cleanup() {
     # dejima-notify-watch's marker consumption.
     rm -rf "${DEJIMA_NOTIFY_DIR:-}" \
         || echo "warning: could not remove the marker directory ${DEJIMA_NOTIFY_DIR:-}" >&2
+    exit_step "session ended - keeping what it compacted away"
+    # Keep whatever the session compacted away, into a store nothing mounts.
+    # `|| true` for the same reason every other line in this trap has it: a
+    # failure here must not replace the container's exit status.
+    #
+    # Ahead of the audit rather than after it, so the audit keeps the last
+    # word (see below): this prints a line of its own whenever it keeps
+    # something, and that line was landing under the verdict.
+    "$SCRIPT_DIR/dejima-epitaph" save "$WORKSPACE_DIR" || true
     # Last writer, structurally: the operator reads bottom-up from their
     # shell prompt, so anything printed after the verdict pushes it out of
     # view. The marker-directory warning above is the one line a session can
@@ -694,6 +843,7 @@ cleanup() {
     # STATE is dropped from the rm -f above and freed only after compare has
     # read it, so this function's own cleanup can't race the audit that
     # needs it - and its removal is guarded too, so it stays truly last.
+    exit_step "checking what changed on the host while it ran"
     DEJIMA_RW_MOUNTS="$DEJIMA_RW_MOUNTS" "$SCRIPT_DIR/dejima-audit" compare "$WORKSPACE_DIR" "$SANDBOX_AUDIT_STATE" || true
     rm -f "$SANDBOX_AUDIT_STATE" || true
 }
@@ -770,6 +920,42 @@ case "$SANDBOX_CMD" in
         ;;
 esac
 export SANDBOX_CMD
+
+# What the wrapper inside the container needs to shadow the harness's binary
+# and pin its flags, resolved here because the manifests are host-side data
+# the container never sees. Every value is the selected harness's own; none
+# of it is written down in the image.
+#
+# SANDBOX_HARNESS_COMMAND is the name to shadow, and is the manifest's rather
+# than [core] command: overriding the latter to `nvim` changes what dejima
+# launches, not which harness is wrapped, and an editor-driven session that
+# then runs the harness from a terminal must still get the flags.
+SANDBOX_HARNESS_COMMAND="$(manifest_get harness "$DEJIMA_HARNESS_NAME" command)"
+SANDBOX_HARNESS_REAL="$(manifest_get harness "$DEJIMA_HARNESS_NAME" command_path)"
+SANDBOX_HARNESS_ARGS="$(manifest_get harness "$DEJIMA_HARNESS_NAME" no_confirm_args)"
+# Named in full rather than templated: the settings layer's container-side
+# path is dejima's own fixed contract (docker-compose.yml mounts it there),
+# not a value that varies per run, and {{...}} in a manifest belongs to the
+# briefing composer's substitution rather than to a second mechanism that
+# merely looks like it.
+SANDBOX_HARNESS_SETTINGS="$(manifest_get harness "$DEJIMA_HARNESS_NAME" settings_args)"
+# Unattended operation is R1 of the capability bar, and a harness whose
+# manifest names no way to ask for it would run asking for confirmation on
+# every tool call, against a container built on the premise that it does not.
+# That fails silently and looks like the harness hanging, so refuse instead.
+if [ -z "$SANDBOX_HARNESS_ARGS" ]; then
+    echo "Harness '$DEJIMA_HARNESS_NAME' declares no no_confirm_args." >&2
+    echo "dejima runs unattended; a harness that stops to ask cannot be run that" >&2
+    echo "way. See \"What a harness must be able to do\" in docs/development.md." >&2
+    exit 1
+fi
+export SANDBOX_HARNESS_COMMAND SANDBOX_HARNESS_REAL SANDBOX_HARNESS_ARGS \
+    SANDBOX_HARNESS_SETTINGS
+
+# [core] command defaults to the selected harness's own command rather than
+# to a literal in core's schema, which is where `claude` used to sit.
+[ -n "$SANDBOX_CMD" ] || SANDBOX_CMD="$SANDBOX_HARNESS_COMMAND"
+export SANDBOX_CMD
 # docker-compose.yml/entrypoint.sh compare these against "1"/"0", not
 # "on"/"off" - config_get_bool is what produces that, config_get would pass
 # "on"/"off" straight through and (among other things) silently disable the
@@ -839,6 +1025,13 @@ for mounted_dir in hooks skills commands agents output-styles; do
         case " $mounted_dirs " in
             *" $real_dir "*) continue ;;
         esac
+        # The target is discovered, not configured, so nothing has vetted it
+        # (#15). A hook symlinked into the approval store would mount it
+        # here, read-only and silently. Refusing rather than skipping: a
+        # skipped mount leaves the symlink dangling inside the container and
+        # the hook failing for a reason nothing states.
+        _guarded_stores
+        _refuse_guarded_mount "~/.claude symlink target" "$real_dir" "${GUARDED_STORES[@]}"
         mounted_dirs="$mounted_dirs $real_dir"
         extra_mounts+=(-v "$real_dir:$real_dir:ro")
     done
@@ -883,9 +1076,15 @@ if [ -n "$SANDBOX_EXTRA_MOUNTS" ]; then
     for path in "${extra_paths[@]}"; do
         [ -n "$path" ] || continue
         if [ ! -e "$path" ]; then
-            echo "[mounts] readonly entry does not exist: $path" >&2
+            echo "$(_mount_label "$path" "[mounts] readonly entry" "--extra-readonly path") does not exist: $path" >&2
             exit 1
         fi
+        # [mounts] read_write has been guarded since it was added; this one
+        # never was, so `readonly = ~/.config/dejima` mounted the approval
+        # store into the container and nothing said a word (#15).
+        _guarded_stores
+        _refuse_guarded_mount "$(_mount_label "$path" "[mounts] readonly entry" "--extra-readonly path")" \
+            "$path" "${GUARDED_STORES[@]}"
         extra_mounts+=(-v "$path:$path:ro")
     done
 fi
@@ -927,36 +1126,30 @@ if [ -d "$WORKSPACE_DIR/.git/hooks" ]; then
     extra_mounts+=(-v "$WORKSPACE_DIR/.git/hooks:$WORKSPACE_DIR/.git/hooks:ro")
 fi
 
-# ~/.claude/statusline.sh is the same class of file as the config staged at
-# /etc/dejima-claude by docker-compose.yml: settings.json's statusLine.command
-# points at it, and Claude Code executes it on the host. Staged the same way,
-# for entrypoint.sh to copy into place, rather than mounted at its real path.
+# The state this harness keeps on the host, and which of it this workspace
+# is entitled to see. Harness-specific by nature - which directories exist,
+# and what decides that one of their entries belongs to this workspace, is a
+# fact about the harness rather than about dejima - so it lives in the
+# harness plugin, which declares each mount through emit_mount. The guarded-
+# store refusal runs there, at the point of declaration, rather than being
+# something each plugin has to remember. See bin/dejima-harness-lib.sh.
+harness_hook harness_state_mounts "$WORKSPACE_DIR"
+
+# statusline.sh is executed on the host by Claude Code (settings.json's
+# statusLine.command), so it is staged for entrypoint.sh to copy rather than
+# mounted at its real path - same as the config in docker-compose.yml.
 #
-# It used to be mounted read-only at $HOME/.claude/statusline.sh, and that
-# protected only the plain-regular-file case - confirmed against a live
-# container, not just reasoned about. When statusline.sh was itself a symlink
-# (the common case - typically into a dotfiles checkout), the read-write
-# ~/.claude mount was established first, so the destination already had that
-# symlink sitting on it, and bind-mount destination resolution (runc's
-# SecureJoin) followed it: the :ro mount landed on the target directory
-# instead of the symlink dentry, which stayed writable and replaceable. There
-# was no mount-based fix while ~/.claude was mounted whole, and the
-# compensating control was detection after the fact (bin/dejima-audit hashes
-# the symlink's target string). Staging closes it outright: Docker resolves
-# the symlink when it sets up the staging mount, so entrypoint.sh copies the
-# real content to a path nothing on the host can reach.
+# A :ro mount there protected only the plain-file case, confirmed against a
+# live container. When statusline.sh is itself a symlink (the common case),
+# bind-mount destination resolution follows it, so the mount lands on the
+# target and the link stays replaceable. Staging closes that: Docker resolves
+# the symlink setting up the staging mount, and the copy lands where the host
+# cannot reach it.
 #
-# Conditional on existence because not every host has one, and mounting a
-# nonexistent source would make Docker create a root-owned directory at the
-# target instead - worse than the gap this closes.
-#
-# A *dangling* symlink (target doesn't exist, e.g. a dotfiles checkout that
-# isn't present) is neither "exists" nor safely mountable, so it gets its own
-# branch that refuses to launch rather than either silently skipping the
-# mount (reopens the gap plain -e had) or attempting it anyway (Docker can't
-# bind-mount a dangling symlink - it fails deep in mount setup with an opaque
-# "mkdir ...: file exists", since the symlink already occupies that path,
-# with nothing connecting that error back to the statusline for the user).
+# Conditional on existence - mounting a missing source makes Docker create a
+# root-owned directory at the target. A dangling symlink is neither
+# mountable nor "exists", so it refuses rather than skipping the mount
+# (reopening the gap) or failing later with an opaque "mkdir: file exists".
 if [ -L "$HOME/.claude/statusline.sh" ] && [ ! -e "$HOME/.claude/statusline.sh" ]; then
     echo "$HOME/.claude/statusline.sh is a symlink to a target that doesn't" >&2
     echo "exist. Docker can't bind-mount a dangling symlink, and skipping the" >&2
@@ -967,16 +1160,12 @@ elif [ -e "$HOME/.claude/statusline.sh" ]; then
     extra_mounts+=(-v "$HOME/.claude/statusline.sh:/etc/dejima-claude/statusline.sh:ro")
 fi
 
-# Keep the real Anthropic OAuth credential out of the container entirely:
-# run a host-side proxy that injects it per-request (see
-# bin/dejima-anthropic-proxy and docs/security.md). Nothing needs masking
-# any more - ~/.claude isn't mounted as a whole, so the real credential (and
-# the proxy's .credentials-proxy-tmp staging directory beside it) simply
-# never crosses into the container. That masking is what #50 was: both masks
-# were single-file/-directory mounts nested in the read-write ~/.claude
-# mount, and the proxy's own write-back renamed the credential out from
-# under them, detaching the mask and exposing the real token from the first
-# refresh on.
+# The real Anthropic OAuth credential never enters the container: a host-side
+# proxy injects it per request (bin/dejima-anthropic-proxy, docs/security.md).
+# Nothing needs masking now that ~/.claude isn't mounted whole - and masking
+# is what #50 was, a single-file mount nested in the read-write ~/.claude
+# that the proxy's own write-back renamed out from under, exposing the real
+# token on the first refresh.
 #
 # How the container then reaches the proxy depends on the transport. In the
 # default tls one, entrypoint.sh points api.anthropic.com at an in-container
@@ -1186,26 +1375,42 @@ if remote_enabled; then
     # own ID.
     extra_mounts+=(-e "DEJIMA_RUN_ID=$DEJIMA_RUN_ID")
 
-    # The notify hook is registered per session, into the override layer from
-    # above and nowhere else: registering it in the operator's real
-    # settings.json would run it for every sandbox and for the host's own
-    # Claude Code, which share that file, when it is only meaningful for a
-    # session that has a marker directory. $SANDBOX_SETTINGS_FILE is
-    # already a throwaway mktemp file, so this rewrite cannot reach the
-    # operator's real settings.json; it is mounted read-only, so the
-    # session cannot unregister itself either. Claude Code merges hooks
-    # across settings layers rather than letting the higher one replace the
-    # lower, so these are added to whatever the operator already has rather
-    # than displacing it. $HOME is left unexpanded for Claude Code to resolve
-    # at hook-run time - inside the container, not here - matching the
-    # convention install.sh already uses.
-    SANDBOX_REMOTE_SETTINGS_TMP="$(mktemp "${TMPDIR:-/tmp}/dejima-settings-remote.XXXXXX.json")"
-    jq '.hooks.Notification = ((.hooks.Notification // []) + [{"matcher": "", "hooks": [{"type": "command", "command": "$HOME/.claude/hooks/dejima-notify.sh waiting"}]}])
-      | .hooks.Stop = ((.hooks.Stop // []) + [{"matcher": "", "hooks": [{"type": "command", "command": "$HOME/.claude/hooks/dejima-notify.sh done"}]}])' \
-        "$SANDBOX_SETTINGS_FILE" > "$SANDBOX_REMOTE_SETTINGS_TMP"
-    mv "$SANDBOX_REMOTE_SETTINGS_TMP" "$SANDBOX_SETTINGS_FILE"
+    # Registered per session, into the override layer from above and nowhere
+    # else: putting it in the operator's real settings.json would run it for
+    # every sandbox and for the host's own agent, which share that file, when
+    # it is only meaningful for a session that has a marker directory.
+    # $SANDBOX_SETTINGS_FILE is already a throwaway mktemp file, so this
+    # rewrite cannot reach the operator's real settings; it is mounted
+    # read-only, so the session cannot unregister itself either.
+    #
+    # How a harness is told to run something on those two events is the
+    # harness's own business - Claude Code merges hook entries across
+    # settings layers, another harness may take a config key or an
+    # environment variable - so the writing is the plugin's. Whether it
+    # happens at all is core's, and is gated on the harness having declared
+    # the notify guard.
+    if harness_has_guard notify; then
+        harness_hook harness_register_notify "$SANDBOX_SETTINGS_FILE"
+    fi
 fi
 
+# Everything from here to the container's own first output is silent: the
+# audit snapshot, an image build that prints nothing when every layer is
+# cached, and Docker creating the container with a mount per allowed path.
+# Long enough to read as hung, so say what is happening first. A first run,
+# or one after a Dockerfile change, builds the image and takes minutes rather
+# than seconds - the line names that, since it is the case where someone is
+# most likely to give up and interrupt.
+printf '%sdejima%s is starting, please wait...\n' "$C_KEY" "$C_RESET" >&2
+printf '%sA first run builds the container image, which takes several minutes.%s\n' \
+    "$C_DIM" "$C_RESET" >&2
+
+# Named here, once, at the moment the operator is choosing to proceed. Silent
+# when the harness hosts every guard, so this is not a line anyone learns to
+# scroll past - it appears only when a protection they have elsewhere is
+# genuinely not in force here.
+harness_report_guards "$DEJIMA_HARNESS_NAME"
+
 # Snapshot host-persistence surfaces so the post-run audit can report
 # anything that changed while this session ran (advisory - see bin/dejima-audit).
 "$SCRIPT_DIR/dejima-audit" snapshot "$WORKSPACE_DIR" "$SANDBOX_AUDIT_STATE" || true
@@ -1259,6 +1464,22 @@ if [ -n "$PROJECT_STORE" ] && [ -s "$PROJECT_STORE/dependencies" ]; then
 fi
 export SANDBOX_IMAGE
 
+# nvim's data, state and cache are Docker volumes rather than host paths, so
+# nothing a session installs reaches the host's own nvim. They did outlive
+# the container under one fixed name each, though, shared by every project
+# on the machine: a session that drops a file under `lazy/`, or replaces a
+# binary under `mason/bin/`, is writing something a later session in an
+# unrelated project executes the next time nvim runs in the sandbox. That is
+# the persistence this repository closes everywhere else, so it is scoped
+# the same way the derived image tag is - per project.
+#
+# The prefix is not decoration: project_slug maps a leading `/` to `-`, and
+# a Docker volume name may not start with one.
+_nvim_volume_slug="$(project_slug "$WORKSPACE_DIR")"
+export SANDBOX_NVIM_DATA="dejima_nvim-data${_nvim_volume_slug}"
+export SANDBOX_NVIM_STATE="dejima_nvim-state${_nvim_volume_slug}"
+export SANDBOX_NVIM_CACHE="dejima_nvim-cache${_nvim_volume_slug}"
+
 # Not exec'd (the EXIT trap cleans up the generated temp files) and not a
 # bare command under set -e (a nonzero/interrupted container run - e.g.
 # Ctrl-C returning 130 - would otherwise abort the script here, skipping the
@@ -1270,7 +1491,7 @@ export SANDBOX_IMAGE
 # because exit-time output can be ignored exactly once, permanently. This is
 # host-side and stays host-side: it must never be routed into the session's
 # context, which would make it a cross-session persistence channel.
-audit_slug="$(printf '%s' "$WORKSPACE_DIR" | sed 's/[^a-zA-Z0-9]/-/g')"
+audit_slug="$(ws_slug "$WORKSPACE_DIR")"
 audit_reports="$AUDIT_DIR/$audit_slug"
 # `|| true`: a workspace with no history yet (the common case - nothing has
 # ever run against it, or every past finding is already acknowledged) has no

bin/dejima-attach

@@ -24,23 +24,17 @@ fi
 STATE_DIR="$(remote_state_dir)"
 SESSIONS_DIR="$STATE_DIR/sessions"
 
-# gruvbox dark. Every value below is upstream's, tagged with the name of the
-# color.table row it came from:
+# gruvbox dark, every value upstream's and tagged with its color.table row:
 # https://github.com/gruvbox-community/gruvbox-contrib/blob/master/color.table
-# Carrying the name and not just the hex is the point - a palette this widely
-# known is one almost nobody has memorised to six digits, so a wrong value is
-# only catchable by reading if there is something to check it against.
+# The names are carried so a wrong hex is catchable by reading.
 #
-# Truecolor only, and deliberately not paired with a 256-colour table: a
-# terminal without it - or tmux without `Tc` - approximates 38;2 to its own
-# nearest colour rather than printing the escape, so the fallback already
-# exists one layer down and a second palette here would only duplicate it.
+# Truecolor only: a terminal without it, or tmux without `Tc`, approximates
+# 38;2 to its nearest colour rather than printing the escape, so the fallback
+# already exists one layer down.
 #
-# Everything below is empty unless stdout is a terminal that says it can show
-# colour, which is what keeps --list clean: it is a machine interface (ttyd's
-# --url-arg feeds run ids back in, and the README documents the rows as
-# scriptable), and it is also how the bats suite reads this script, so its
-# assertions see the same bare words either way.
+# Empty unless stdout is a colour-capable terminal, which keeps --list a
+# machine interface (ttyd feeds run ids back in) and is also how the bats
+# suite reads this script.
 if [ -t 1 ] && [ -z "${NO_COLOR:-}" ] && [ "${TERM:-dumb}" != dumb ]; then
     C_OFF=$'\033[0m'; C_BOLD=$'\033[1m'
     C_BADGE=$'\033[38;2;254;128;25m'    # bright_orange #fe8019, the wordmark
@@ -95,22 +89,16 @@ state_color() {
 # How long after its own started= a registered session may still have no
 # container, before the prune in list_sessions treats that as "gone".
 #
-# Sized against the step that actually holds a launch up: bin/dejima writes
-# the registry entry, then rewrites the settings copy, snapshots the
-# workspace for the audit, runs an UNCONDITIONAL `docker compose build
-# claude`, optionally builds up to two derived images on top of it, and only
-# then reaches `docker compose run`. A cold base build is the Dockerfile's
-# apt and npm layers fetched over the network - minutes, not seconds - and
-# the derived builds add to it, while a warm run reaches the container in
-# seconds and never comes near this window.
+# Sized against what holds a launch up: the registry entry is written before
+# an unconditional `docker compose build`, plus up to two derived image
+# builds, and a cold base build fetches apt and npm layers - minutes. A warm
+# run never comes near this window.
 #
-# The two errors are not symmetric, which is why this is generous. Too long
-# only leaves a session that really did die listed until the window passes,
-# and the next listing prunes it. Too short deletes the registry entry of a
-# session that is about to start: bin/dejima never rewrites that entry, so
-# the session is permanently absent from both front doors, and the marker
-# directory deleted with it is the source of a bind mount the imminent
-# `docker compose run` declares.
+# Generous because the errors are not symmetric. Too long leaves a dead
+# session listed until the window passes. Too short deletes the entry of a
+# session about to start: bin/dejima never rewrites it, so the session is
+# permanently absent from both front doors, and the marker directory removed
+# with it is a bind source the imminent `compose run` declares.
 STARTUP_GRACE=900
 
 # Parsed, never sourced: this file is host-written, but treating state files
@@ -281,26 +269,21 @@ attach() {
         exit 1
     fi
     group="dejima-$run_id"
-    # Both values come from the same live pane, in one round trip. Taking the
-    # session name from the recorded label instead would follow a name rather
-    # than the sandbox: an operator renaming sessions is enough to point it at
-    # an unrelated session that has since taken the old name.
+    # Both values from the same live pane, one round trip. Taking the session
+    # name from the recorded label would follow a name rather than the
+    # sandbox, and a rename is enough to point it at an unrelated session.
     #
-    # The name is a starting point, never a guarantee. It is frequently
-    # ANOTHER remote client's throwaway group (tmux reports whichever session
-    # holding the window is most recent, and every grouped session here is
-    # newer than the operator's own), and those carry destroy-unattached - so
-    # the base can be made to stop existing on demand by anyone holding a
-    # second connection. That is why the session created from it is verified
-    # below instead of trusted: `new-session -t` does not fail on a name that
-    # matches nothing, it silently creates a fresh UNGROUPED session, which
-    # for a remote client is a plain host shell with no sandbox in it.
+    # The name is a starting point, not a guarantee: it is often another
+    # remote client's throwaway group, and those carry destroy-unattached, so
+    # anyone with a second connection can make the base stop existing. Hence
+    # the verification below - `new-session -t` does not fail on a name
+    # matching nothing, it silently creates an UNGROUPED session, which for a
+    # remote client is a plain host shell with no sandbox in it.
     #
-    # display-message prints nothing and still exits 0 for a target that no
-    # longer exists, so an empty result means the pane went away between the
-    # check above and here. That case is caught by the verification below too;
-    # bailing early just avoids creating a junk session to immediately kill,
-    # and says what actually happened.
+    # display-message prints nothing and exits 0 for a vanished target, so an
+    # empty result means the pane went away since the check above. The
+    # verification catches that too; bailing early just avoids creating a
+    # junk session and says what happened.
     target="$(tmux display-message -p -t "$entry_pane" '#{session_name}:#{window_index}' 2>/dev/null || true)"
     # Session names may contain ':', window indexes can't.
     base="${target%:*}"

bin/dejima-audit

@@ -12,6 +12,10 @@ set -uo pipefail
 source "$(dirname "$(readlink -f "$0")")/dejima-provider-lib.sh"
 # shellcheck source=dejima-color-lib.sh
 source "$(dirname "$(readlink -f "$0")")/dejima-color-lib.sh"
+# shellcheck source=dejima-slug-lib.sh
+source "$(dirname "$(readlink -f "$0")")/dejima-slug-lib.sh"
+# shellcheck source=dejima-harness-lib.sh
+source "$(dirname "$(readlink -f "$0")")/dejima-harness-lib.sh"
 # Defined empty here so every interpolation below is safe before a surface
 # picks its stream; compare (stderr) and ack (stdout) each re-init for theirs.
 color_init 1
@@ -19,17 +23,36 @@ config_validate || exit 1
 
 cmd="${1:-}"; ws_arg="${2:-}"; state="${3:-}"
 
-PROJECTS_DIR="$(config_get harness.claude-code projects_dir)"
+PROJECTS_DIR="$(harness_config_get projects_dir)"
 
-# The one top-level ~/.claude entry docker-compose.yml's read-only overlays
-# don't cover (settings.json/CLAUDE.md/plugins/hooks/skills/commands/agents/
-# output-styles all are) - settings.json's statusLine.command runs it on the
-# HOST, unsandboxed, and bin/dejima's own read-only mount onto it
-# doesn't reliably protect it when it's a symlink (the common case - runc
-# resolves the mount destination through the symlink; confirmed against a
-# live container, see test/runtime.bats). Detection here is the compensating
-# control for that gap. Overridable for tests.
-STATUSLINE_PATH="$(config_get harness.claude-code statusline_path)"
+# The paths the selected harness causes to run ON THE HOST, unsandboxed,
+# from configuration a session can reach - Claude Code's statusline is the
+# instance that exists. Each is watched on its own rather than folded into
+# hash_paths, because bin/dejima's read-only mount does not reliably protect
+# one that is a symlink (the common case - runc resolves the mount
+# destination through the link; confirmed against a live container, see
+# test/runtime.bats), so detection here is the compensating control.
+#
+# Declared by the harness's plugin, not by its manifest: the labels below
+# are the fixed prefixes `ack` strips to recover an id, which makes them
+# part of a parser's grammar rather than data. See bin/dejima-harness-lib.sh.
+harness_plugin_load "$(config_get core harness)"
+host_exec_load
+
+# With no harness selected there is no statusline path to watch, and the
+# compensating control above quietly covers nothing. bin/dejima refuses to
+# start without [core] harness, so a snapshot in this state means the audit
+# was driven by hand against a config dejima itself would reject - say so
+# rather than reporting a clean run over a surface that was never examined.
+# A silently narrower audit is worse than a noisy one.
+if [ "${#DEJIMA_HOST_EXEC[@]}" -eq 0 ] && [ -z "$(config_get core harness)" ]; then
+    case "$cmd" in
+        snapshot|compare)
+            echo "dejima audit: no [core] harness is set, so harness-specific surfaces" >&2
+            echo "              (whatever this harness runs on the host) are NOT checked." >&2
+            ;;
+    esac
+fi
 
 # This script's own repo root - the sandbox's trusted computing base (see
 # tcb_snapshot below) is only monitored when the workspace being audited IS
@@ -53,7 +76,9 @@ AUDIT_DIR="$(config_get audit dir)"
 # suggesting a restart is an entirely ordinary thing to suggest.
 AUDIT_KEEP="$(config_get audit keep)"
 
-ws_slug() { printf '%s' "$1" | sed 's/[^a-zA-Z0-9]/-/g'; }
+# ws_slug and workspace_slugs come from dejima-slug-lib.sh, shared with
+# bin/dejima so the set of slugs it mounts and the set treated as this
+# workspace's own here cannot drift apart.
 
 # The one report both `show` and `ack` operate on: the OLDEST unacknowledged
 # CRITICAL report if there is one, else the newest report overall.
@@ -69,6 +94,24 @@ ws_slug() { printf '%s' "$1" | sed 's/[^a-zA-Z0-9]/-/g'; }
 # startup warning may well reach for `show` first, and showing them a
 # different report than the one being warned about sends them looking for a
 # finding that is not in it.
+# How many reports are waiting, and how many of those are CRITICAL. `ack`
+# walks one report at a time by design - each is a separate session's worth of
+# findings - but saying nothing about the rest is what made clearing a backlog
+# feel like the tool was stuck on one.
+# A report's stable identity: its basename without the -CRITICAL marker, which
+# `ack` removes once every CRITICAL in it has been dealt with.
+report_key() {
+    local b; b="$(basename "$1")"
+    printf '%s' "${b%-CRITICAL.txt}.txt" | sed 's/\.txt\.txt$/.txt/'
+}
+
+count_reports() {
+    local dir="$1"
+    printf '%s %s\n' \
+        "$(find "$dir" -maxdepth 1 -name '*.txt' 2>/dev/null | wc -l)" \
+        "$(find "$dir" -maxdepth 1 -name '*-CRITICAL.txt' 2>/dev/null | wc -l)"
+}
+
 select_report() {
     local dir="$1" r
     r="$(find "$dir" -maxdepth 1 -name '*-CRITICAL.txt' 2>/dev/null | sort | head -1)"
@@ -83,19 +126,16 @@ BASELINE_DIR="$(config_get audit baselines_dir)"
 # should mean that repo. bin/dejima always passes an explicit path, so this
 # only ever fires for a hand-run show/ack.
 #
-# Candidates in order, first one that already HAS reports wins - existence of
-# a report directory is the only evidence available about which path
-# bin/dejima was actually invoked with, and the three candidates can
-# genuinely differ:
-#   1. the enclosing repo's toplevel (the ordinary case, and from a subdir);
-#   2. the MAIN worktree root, when cwd is a linked worktree - the sandbox's
-#      own .worktrees/<branch> workflow keys reports to the parent repo;
-#   3. $PWD, which covers a non-git directory and the case where the path
-#      bin/dejima was given (logical, via `cd && pwd`) differs from git's
-#      physical toplevel because of a symlinked path component.
-# Falling back to candidate 1 when none of them has reports is purely so the
-# "no reports for X" message can name a concrete path: a wrong guess must be
-# visible rather than silent.
+# Candidates in order, first one that already HAS reports wins - a report
+# directory is the only evidence of which path bin/dejima was invoked with,
+# and the three can genuinely differ:
+#   1. the enclosing repo's toplevel (ordinary case, and from a subdir);
+#   2. the MAIN worktree root, when cwd is a linked worktree - the
+#      .worktrees/<branch> workflow keys reports to the parent repo;
+#   3. $PWD, for a non-git directory, or when the logical path bin/dejima was
+#      given differs from git's physical toplevel through a symlink.
+# Falling back to 1 when none has reports is so "no reports for X" can name a
+# concrete path: a wrong guess must be visible rather than silent.
 resolve_ws() {
     local given="$1" top common main c
     if [ -n "$given" ]; then
@@ -350,11 +390,14 @@ withhold() { printf '%s\n' "$2" > "$1" 2>/dev/null || true; }
 # compare against.
 existed_before() {
     local rid="$1" range enc
+    if host_exec_is_id "$rid"; then
+        range='/^### hostexec/,$p'; enc="$(enc_id "${rid#*:}")"
+    else
     case "$rid" in
-        statusline:*) range='/^### statusline/,$p';           enc="$(enc_id "${rid#statusline:}")" ;;
-        tcb:*)        range='/^### tcb/,/^### statusline/p';   enc="$(enc_id "${rid#tcb:}")" ;;
+        tcb:*)        range='/^### tcb/,/^### hostexec/p';    enc="$(enc_id "${rid#tcb:}")" ;;
         *)            range='/^### paths/,/^### projects/p';   enc="$(enc_id "$rid")" ;;
     esac
+    fi
     sed -n "$range" "$state" | grep -v '^### ' | cut -f1 | grep -qxF -- "$enc"
 }
 
@@ -499,11 +542,15 @@ promote_finding() {
     # Mirrors current_content_for_id's dispatch; gitconfig ids never reach
     # here (promote_gitconfig returned above), and stat does not follow a
     # symlink, whose rendering is its target string rather than its content.
+    # A host-execution surface's path comes from the harness plugin that
+    # declared it, not from the workspace.
+    local _hp
+    if _hp="$(host_exec_path_for_id "$rid")"; then src="$_hp"; else
     case "$rid" in
-        statusline:*) src="$STATUSLINE_PATH" ;;
         tcb:*)        src="$ws/${rid#tcb:}" ;;
         *)            src="$ws/$rid" ;;
     esac
+    fi
     csz="$(stat -c %s "$src" 2>/dev/null)" || csz=0
     if [ "$csz" -gt "$STAGE_FILE_CAP" ]; then
         withhold "$sc/$key.withheld" "change too large to retain - compare by hand"
@@ -568,6 +615,11 @@ baseline_suppresses() {
 }
 
 # Whether `ack` has already walked this report to its end. Recorded in the
+# Recorded against the report's name with any -CRITICAL suffix stripped,
+# because a report that sheds that suffix later in the same run would
+# otherwise carry a marker naming a file that no longer exists - and re-prompt
+# forever, which is the failure the marker exists to prevent.
+#
 # baseline file, alongside the approvals, as `reviewed <report-basename>`:
 # the file is already the per-workspace record of what the operator has
 # decided, already parsed line-by-line, and already tolerant of verbs it does
@@ -637,7 +689,7 @@ id_severity() {
         gitconfig:*) k="${1#gitconfig:}"; k="${k#worktree:}"
                      is_exec_key "$k" && { echo CRITICAL; return; }
                      case "$k" in alias.*) echo CRITICAL ;; *) echo WARNING ;; esac ;;
-        statusline:*|tcb:*) echo CRITICAL ;;
+        tcb:*) echo CRITICAL ;;
         *) path_severity "$1" ;;
     esac
 }
@@ -654,37 +706,26 @@ sanitize() {
         | LC_ALL=C sed $'s/\xc2[\x80-\x9f]//g; s/\xe2\x80[\xaa-\xae]//g; s/\xe2\x81[\xa6-\xa9]//g'
 }
 
-# sanitize() is lossy, so it must never be what an ID passes through.
+# sanitize() is lossy, so an ID must never pass through it. `ack` derives the
+# baseline id from the report text and the next `compare` matches it against
+# the RAW bytes on disk, so a stripped id cannot round-trip: the surface
+# becomes permanently unsuppressible and the startup warning fires forever.
 #
-# `ack` derives the baseline id it records from the report's text, and the
-# next `compare` matches that id against the RAW bytes on disk. An id
-# containing anything sanitize() strips therefore could not round-trip at
-# all: the surface stayed permanently unsuppressible and its report
-# permanently unclearable, so bin/dejima's startup warning fired forever -
-# the alert fatigue this whole feature exists to end, reached from the other
-# side. Ids are encoded instead, and the encoded form is canonical
-# everywhere it is written down: in the report AND in the baseline file.
+# Ids are encoded instead, canonically, in the report AND the baseline file.
+# An id sanitize() leaves alone is written verbatim (every ordinary one, so
+# existing baselines are unchanged); anything else is hex-encoded behind a
+# marker, which is ASCII and reverses exactly.
 #
-# An id that survives sanitize() unchanged is already display-safe and is
-# written verbatim - that is every ordinary id, so reports and existing
-# baseline files are byte-for-byte what they were. Anything else is
-# hex-encoded behind a marker, which is ASCII by construction and reverses
-# exactly.
+# An id that merely LOOKS encoded is encoded too, so the marker's presence
+# always means this encoder put it there. Otherwise a workspace could present
+# an id `ack` decodes into one of the attacker's choosing - baselining
+# something other than the finding while counting it acknowledged. Only the
+# START matters, and no monitored path begins with the marker, but a git
+# config section name is free text the session can write.
 #
-# An id that merely LOOKS encoded is encoded too, so that decoding stays
-# unambiguous: the marker's presence always means this encoder put it there.
-# Without that, a workspace could present an id that `ack` decoded into one
-# of the attacker's choosing - baselining something other than the finding
-# while still counting it acknowledged, the same shape as the id-derivation
-# hijacks the anchored case arms in `ack` guard against. Only the START of
-# an id is significant, and no monitored path can begin with the marker
-# (they are all rooted at .claude/.vscode/.git/.envrc/.mcp.json), but a git
-# config section name is free text in a file the session can write and the
-# audit parses that file directly rather than through git.
-#
-# The encoded form is also the honest rendering. sanitize() DELETES those
-# bytes, so a planted ".git/hooks/pre<U+202E>commit" displayed as
-# "precommit" sends an operator grepping for a name no file on disk has.
+# Encoding is also the honest rendering: sanitize() DELETES those bytes, so a
+# planted ".git/hooks/pre<U+202E>commit" shown as "precommit" sends an
+# operator grepping for a name no file has.
 HEX_MARK='{hex}'
 enc_id() {
     local v="$1"
@@ -787,6 +828,22 @@ hash_paths() {
     } | sort
 }
 
+# Whether `git status` would show a change to $2 inside repo $1. One call per
+# candidate path rather than one listing compared as a set: the candidate list
+# is what changed this session, which is small, and --porcelain on a path is
+# cheap. A path git does not know about at all (never added, and ignored)
+# prints nothing, which is exactly the invisible case.
+#
+# A repo git refuses to read (not a repository, or a dubious-ownership
+# refusal) makes this return false, so every change is reported rather than
+# silently downgraded - failing towards the noisier answer, not the quieter
+# one.
+git_shows_change() {
+    local repo="$1" path="$2" out
+    out="$(git -C "$repo" status --porcelain --untracked-files=all -- "$path" 2>/dev/null)" || return 1
+    [ -n "$out" ]
+}
+
 tcb_snapshot() {
     # encoded-path<TAB>sha for the sandbox's own trusted computing base, but
     # only when $ws IS the sandbox's repo (self-mount) - see AUDIT_REPO_ROOT.
@@ -868,13 +925,14 @@ gitconfig_sha_for_key() {
 # that namespace, so the id/sha pair ack records is exactly what
 # baseline_suppresses will later compare against.
 current_sha_for_id() {
-    local id="$1" f
+    local id="$1" f _hp
+    if _hp="$(host_exec_path_for_id "$id")"; then f="$_hp"; else
     case "$id" in
         gitconfig:*)  gitconfig_sha_for_key "$ws" "${id#gitconfig:}"; return ;;
-        statusline:*) f="$STATUSLINE_PATH" ;;
         tcb:*)        f="$ws/${id#tcb:}" ;;
         *)            f="$ws/$id" ;;
     esac
+    fi
     # -e alone follows a symlink and misses one whose target is gone - but a
     # dangling symlink is exactly what hash_entry() below can still hash (the
     # link's target string, never the target's content), so excluding it here
@@ -895,13 +953,14 @@ DISPLAY_LINE_CAP=40
 # as their target, because that is what hash_entry() hashes for them - the
 # link's own content is not what changed.
 current_content_for_id() {
-    local id="$1" f
+    local id="$1" f _hp
+    if _hp="$(host_exec_path_for_id "$id")"; then f="$_hp"; else
     case "$id" in
         gitconfig:*)  gitconfig_value_for_key "$ws" "${id#gitconfig:}" ;;
-        statusline:*) f="$STATUSLINE_PATH" ;;
         tcb:*)        f="$ws/${id#tcb:}" ;;
         *)            f="$ws/$id" ;;
     esac
+    fi
     case "$id" in gitconfig:*) return ;; esac
     [ -e "$f" ] || [ -L "$f" ] || return 1
     if [ -L "$f" ]; then
@@ -970,10 +1029,13 @@ show_capped() {
 # Never sanitized through a path that would make it unfindable: it is the
 # audit's OWN text about a location, not workspace-controlled content.
 source_hint() {
-    local id="$1"
+    local id="$1" _hp
+    if _hp="$(host_exec_path_for_id "$id")"; then
+        printf '%s\n' "    (full text: $(sanitize "$_hp"))"
+        return
+    fi
     case "$id" in
         gitconfig:*)  printf '%s\n' "    (full value: git -C $ws config --get $(sanitize "${id#gitconfig:}"))" ;;
-        statusline:*) printf '%s\n' "    (full text: $STATUSLINE_PATH)" ;;
         tcb:*)        printf '%s\n' "    (full text: $ws/$(sanitize "${id#tcb:}"))" ;;
         *)            printf '%s\n' "    (full text: $ws/$(sanitize "$id"))" ;;
     esac
@@ -1021,34 +1083,33 @@ render_finding() {
     return 0
 }
 
-projects_snapshot() {
-    # slug<TAB>maxmtime for each immediate subdir of PROJECTS_DIR, sorted by
-    # slug; maxmtime is 0 if the slug dir has no files.
-    local dir="$1" d slug mtime
-    [ -d "$dir" ] || return 0
-    for d in "$dir"/*/; do
-        [ -d "$d" ] || continue
-        slug="$(basename "$d")"
-        mtime="$(find "$d" -type f -printf '%T@\n' 2>/dev/null | sort -n | tail -1)"
-        mtime="${mtime%%.*}"
-        printf '%s\t%s\n' "$slug" "${mtime:-0}"
-    done | sort
-}
-
-statusline_snapshot() {
-    # path<TAB>sha for $HOME/.claude/statusline.sh, or nothing if this host
-    # has none - see STATUSLINE_PATH above for why this one path is
-    # monitored on its own rather than as part of hash_paths.
-    local f="$1"
-    [ -e "$f" ] || return 0
-    stage_entry "statusline:statusline.sh" "$f"
-    printf 'statusline.sh\t%s\n' "$(hash_entry "$f")"
+host_exec_snapshot() {
+    # leaf<TAB>sha for every host-execution surface the harness declared, or
+    # nothing where the host has none. One line per surface, keyed by the
+    # leaf so the id a baseline stores ("<namespace>:<leaf>") is unchanged
+    # for a harness already in use.
+    local rec ns leaf path label
+    for rec in ${DEJIMA_HOST_EXEC[@]+"${DEJIMA_HOST_EXEC[@]}"}; do
+        IFS=$'\t' read -r ns leaf path label <<< "$rec"
+        [ -n "$path" ] && [ -e "$path" ] || continue
+        stage_entry "$ns:$leaf" "$path"
+        printf '%s\t%s\n' "$leaf" "$(hash_entry "$path")"
+    done
 }
 
 case "$cmd" in
     --print-paths)
-        printf 'projects_dir=%s\nstatusline_path=%s\naudit_dir=%s\nbaselines_dir=%s\nkeep=%s\n' \
-            "$PROJECTS_DIR" "$STATUSLINE_PATH" "$AUDIT_DIR" "$BASELINE_DIR" "$AUDIT_KEEP"
+        # statusline_path is still named, because it is the surface this
+        # harness declares and existing callers read it by that name. It
+        # comes from the declaration now rather than from a global of its
+        # own, so a harness declaring none prints an empty value and one
+        # declaring several prints a line each.
+        printf 'projects_dir=%s\naudit_dir=%s\nbaselines_dir=%s\nkeep=%s\n' \
+            "$PROJECTS_DIR" "$AUDIT_DIR" "$BASELINE_DIR" "$AUDIT_KEEP"
+        for _rec in ${DEJIMA_HOST_EXEC[@]+"${DEJIMA_HOST_EXEC[@]}"}; do
+            IFS=$'\t' read -r _ns _leaf _path _label <<< "$_rec"
+            printf '%s_path=%s\n' "$_ns" "$_path"
+        done
         exit 0
         ;;
     snapshot)
@@ -1060,9 +1121,13 @@ case "$cmd" in
         { echo "### staging $STAGE_DIR";
           echo "### gitconfig"; git_config_snapshot "$ws";
           echo "### paths";     hash_paths "$ws";
-          echo "### projects";  projects_snapshot "$PROJECTS_DIR";
+          # Empty since the cross-project finding was retired, and kept
+          # rather than dropped: it is a range boundary for the paths section
+          # above, and removing it would make a state file written by one
+          # version unparseable by the other across an upgrade mid-session.
+          echo "### projects";
           echo "### tcb";       tcb_snapshot "$ws";
-          echo "### statusline"; statusline_snapshot "$STATUSLINE_PATH"; } > "$state" 2>/dev/null || true
+          echo "### hostexec"; host_exec_snapshot; } > "$state" 2>/dev/null || true
         exit 0
         ;;
     compare)
@@ -1099,9 +1164,9 @@ case "$cmd" in
 
         new="$(mktemp)"; { echo "### gitconfig"; git_config_snapshot "$ws";
                            echo "### paths";     hash_paths "$ws";
-                           echo "### projects";  projects_snapshot "$PROJECTS_DIR";
+                           echo "### projects";
                            echo "### tcb";       tcb_snapshot "$ws";
-                           echo "### statusline"; statusline_snapshot "$STATUSLINE_PATH"; } > "$new" 2>/dev/null || {
+                           echo "### hostexec"; host_exec_snapshot; } > "$new" 2>/dev/null || {
             echo "dejima audit: could not complete, skipping." >&2; rm -f "$new"; exit 0; }
 
         # One raw namespaced id per finding whose surface has retainable
@@ -1190,47 +1255,82 @@ case "$cmd" in
                          <(sed -n '/^### paths/,/^### projects/p' "$new"   | grep -v '^### ') \
                  | sed 's/^\t//' | cut -f1 | sort -u )
 
-        # projects diffs: a write to another project's ~/.claude/projects
-        # slug during this session (transcript/memory-poisoning surface).
-        # "self" = current workspace's own slug or a worktree child of it.
-        current_slug="$(ws_slug "$ws")"
-        while IFS= read -r line; do
-            slug="${line%%$'\t'*}"
-            case "$slug" in
-                "$current_slug"|"$current_slug"-*) continue ;;
-            esac
-            warn+=("CROSS-PROJECT history/memory write: $(sanitize "$slug")")
-        done < <(comm -13 <(sed -n '/^### projects/,/^### tcb/p' "$state" | grep -v '^### ') \
-                          <(sed -n '/^### projects/,/^### tcb/p' "$new"   | grep -v '^### ') )
+        # There is deliberately no cross-project finding here any more. It
+        # reported a write to another project's ~/.claude/projects slug, and
+        # a session could once make one: the whole tree was mounted. Since
+        # #104 it gets its own slugs and nothing else, so the only writes this
+        # could still see are other host or dejima sessions doing their own
+        # legitimate work - and with several sessions running it reported all
+        # of them, on every report, drowning the findings that mattered.
+        #
+        # Detection that cannot observe the thing it was built for, and fires
+        # constantly on the thing it was not, is worse than no detection: it
+        # trains the operator to skim. The mount is the control now, and
+        # static.bats pins it.
 
         # tcb diffs: any change under the sandbox's own trusted computing
         # base, only populated when this session's workspace IS the
         # sandbox's own repo (see tcb_snapshot) - empty/no-op otherwise.
-        # Bounded at ### statusline (not $p, the last section now) so this
+        # Bounded at ### hostexec (not $p, the last section now) so this
         # doesn't also swallow that section's lines.
+        # Split on whether git can see the change, because that decides
+        # whether this finding is telling the operator anything.
+        #
+        # tcb_snapshot only populates when the workspace IS dejima's own repo,
+        # so every change here is also a change to a git working tree the
+        # operator is about to read in `git status` and in the pull request.
+        # Reporting each one as a CRITICAL to acknowledge duplicated a
+        # stronger control and made ordinary development unusable.
+        #
+        # What git cannot show is the case worth keeping: a file listed in
+        # .gitignore or .git/info/exclude, or marked assume-unchanged or
+        # skip-worktree, is live code that runs on the next launch and appears
+        # nowhere in `git status`. Nothing but this hash sees it, so it stays
+        # a CRITICAL finding.
+        tcb_visible=0
         while IFS= read -r line; do
-            # Encoded on arrival, same as the path loop above.
             rline="$(dec_id "$line")"
             tid="tcb:$rline"; tsha=""; [ -e "$ws/$rline" ] && tsha="$(hash_entry "$ws/$rline")"
             baseline_suppresses "$tid" "$tsha" && continue
-            crit+=("TCB (sandbox self) modified: $line")
+            if git_shows_change "$ws" "$rline"; then
+                tcb_visible=$((tcb_visible + 1))
+                continue
+            fi
+            crit+=("TCB (sandbox self) modified, INVISIBLE to git: $line")
             finding_ids+=("$tid")
-        done < <(comm -3 <(sed -n '/^### tcb/,/^### statusline/p' "$state" | grep -v '^### ') \
-                         <(sed -n '/^### tcb/,/^### statusline/p' "$new"   | grep -v '^### ') \
+        done < <(comm -3 <(sed -n '/^### tcb/,/^### hostexec/p' "$state" | grep -v '^### ') \
+                         <(sed -n '/^### tcb/,/^### hostexec/p' "$new"   | grep -v '^### ') \
                  | sed 's/^\t//' | cut -f1 | sort -u )
+        if [ "$tcb_visible" -gt 0 ]; then
+            info+=("$tcb_visible change(s) under the sandbox's own bin/, hooks/, docker/, test/ - this workspace IS dejima; review them with git")
+        fi
 
-        # statusline diff: ~/.claude/statusline.sh added/changed/removed/
-        # repointed - see STATUSLINE_PATH above for why this gets its own
-        # section instead of folding into hash_paths. Worded at least as
-        # strongly as COMMAND-EXECUTION/TCB above: this is unsandboxed host
-        # code execution, not just a config change.
+        # host-execution diff: a path the harness runs on the HOST added,
+        # changed, removed or repointed. Its own section rather than folded
+        # into hash_paths for the reason given where the surfaces are loaded,
+        # and worded at least as strongly as COMMAND-EXECUTION/TCB above:
+        # this is unsandboxed host code execution, not just a config change.
+        #
+        # The label is the plugin's, because it is the fixed prefix `ack`
+        # strips to recover the id and so belongs in trusted code rather than
+        # in manifest data. A leaf with no declaring surface cannot happen
+        # from a snapshot this run wrote, but it can from a state file
+        # written before the harness changed - such a line is skipped rather
+        # than reported under a label nothing owns.
         while IFS= read -r line; do
-            sid="statusline:$line"; ssha=""; [ -e "$STATUSLINE_PATH" ] && ssha="$(hash_entry "$STATUSLINE_PATH")"
+            _hl=""; _hns=""; _hpath=""
+            for _rec in ${DEJIMA_HOST_EXEC[@]+"${DEJIMA_HOST_EXEC[@]}"}; do
+                IFS=$'\t' read -r _ns _leaf _path _label <<< "$_rec"
+                [ "$_leaf" = "$line" ] || continue
+                _hns="$_ns"; _hpath="$_path"; _hl="$_label"; break
+            done
+            [ -n "$_hl" ] || continue
+            sid="$_hns:$line"; ssha=""; [ -e "$_hpath" ] && ssha="$(hash_entry "$_hpath")"
             baseline_suppresses "$sid" "$ssha" && continue
-            crit+=("HOST-CODE-EXECUTION statusline.sh changed (runs UNSANDBOXED on the host via settings.json statusLine.command): $(enc_id "$line")")
+            crit+=("$_hl$(enc_id "$line")")
             finding_ids+=("$sid")
-        done < <(comm -3 <(sed -n '/^### statusline/,$p' "$state" | grep -v '^### ') \
-                         <(sed -n '/^### statusline/,$p' "$new"   | grep -v '^### ') \
+        done < <(comm -3 <(sed -n '/^### hostexec/,$p' "$state" | grep -v '^### ') \
+                         <(sed -n '/^### hostexec/,$p' "$new"   | grep -v '^### ') \
                  | sed 's/^\t//' | cut -f1 | sort -u )
 
         rm -f "$new"
@@ -1368,15 +1468,41 @@ case "$cmd" in
         exit 0
         ;;
     show)
+        color_init 1
         report_dir="$AUDIT_DIR/$(ws_slug "$ws")"
         report="$(select_report "$report_dir")" \
             || { echo "dejima audit: no reports for $ws" >&2; exit 0; }
-        # Which report, on stderr so stdout stays the report content alone:
-        # `show` no longer necessarily prints the newest one, and an operator
-        # cross-referencing the run timestamps in the startup warning needs to
-        # know which of them they are looking at.
-        echo "dejima audit: $report" >&2
-        cat "$report"
+        # Default is an index, not a dump. `cat` of one report was a wall of
+        # text that answered neither question an operator actually has - how
+        # much is waiting, and which of it still needs a decision - and it
+        # printed findings already acknowledged, so the wall did not even
+        # shrink as they worked through it.
+        if [ "${2:-}" = --full ] || [ "${3:-}" = --full ] || [ "${4:-}" = --full ]; then
+            echo "dejima audit: $report" >&2
+            cat "$report"
+            exit 0
+        fi
+        read -r n_all n_crit <<< "$(count_reports "$report_dir")"
+        printf '%s report(s) for %s' "$n_all" "$ws"
+        [ "$n_crit" -gt 0 ] && printf ', %s%s with unacknowledged CRITICAL%s' \
+            "$C_CRIT" "$n_crit" "$C_RESET"
+        printf '\n\n'
+        # Newest last, so the one `ack` will offer first is the one nearest
+        # the prompt - and mark it, since select_report prefers CRITICAL over
+        # recency and "the newest" is the wrong guess.
+        r=""; line=""; pending=""
+        while IFS= read -r r; do
+            pending=$(grep -cE '^(CRITICAL|WARNING|INFO)' "$r" 2>/dev/null || true)
+            line="$(printf '  %-44s %s finding(s)' "$(basename "$r")" "${pending:-0}")"
+            case "$r" in
+                *-CRITICAL.txt) printf '%s%s%s' "$C_CRIT" "$line" "$C_RESET" ;;
+                *)              printf '%s' "$line" ;;
+            esac
+            [ "$r" = "$report" ] && printf '   %s<- ack starts here%s' "$C_KEY" "$C_RESET"
+            printf '\n'
+        done < <(find "$report_dir" -maxdepth 1 -name '*.txt' 2>/dev/null | sort)
+        printf '\n%sdejima-audit ack%s walks them; %sdejima-audit show --full%s prints one.\n' \
+            "$C_KEY" "$C_RESET" "$C_KEY" "$C_RESET"
         exit 0
         ;;
     ack)
@@ -1385,6 +1511,17 @@ case "$cmd" in
         bf="$BASELINE_DIR/$(ws_slug "$ws")"
         report="$(select_report "$report_dir")" \
             || { echo "dejima audit: no reports for $ws" >&2; exit 0; }
+        read -r n_all n_crit <<< "$(count_reports "$report_dir")"
+        if [ "$n_all" -gt 1 ]; then
+            printf '%s reports waiting' "$n_all"
+            [ "$n_crit" -gt 0 ] && printf ', %s%s with unacknowledged CRITICAL%s' \
+                "$C_CRIT" "$n_crit" "$C_RESET"
+            printf '.\n\n'
+        fi
+        # One report per pass, every report in one invocation. Walking one and
+        # exiting meant clearing a backlog took as many runs as there were
+        # reports, with nothing on screen to say another was behind it.
+        while :; do
         sc="$(sidecar_dir "$report")"
         # Removed only at the end, and only for findings actually
         # acknowledged - an abort must leave the operator exactly where they
@@ -1403,6 +1540,10 @@ case "$cmd" in
         # (or a non-path CRITICAL, which can never be "approved" - see
         # ackable below) must keep its suffix and stay eviction-protected.
         crit_seen=0; crit_acked=0
+        # Set when a finding was cleared because its surface no longer exists.
+        # Like unackable_shown, it means the report has been dealt with as far
+        # as it can be, so it earns a `reviewed` marker.
+        removed_cleared=0
         # Findings passed over because the baseline already covers them, or
         # because this report has been walked before - counted so a run that
         # displays nothing can say why instead of looking broken.
@@ -1412,7 +1553,7 @@ case "$cmd" in
         # end. Together they decide if a `reviewed` line gets written below.
         unackable_shown=0
         report_reviewed=0
-        if baseline_reviewed "$(basename "$report")"; then report_reviewed=1; fi
+        if baseline_reviewed "$(report_key "$report")"; then report_reviewed=1; fi
         # The loop below reads report lines on its own stdin (`< "$report"`),
         # so operator keystrokes are read from fd 3 - a copy of the script's
         # original stdin taken before that redirection - rather than
@@ -1440,48 +1581,50 @@ case "$cmd" in
                 *) continue ;;
             esac
             [ "$tier" = CRITICAL ] && crit_seen=$((crit_seen + 1))
-            # Derive the namespaced id baseline_suppresses() will later
-            # compare against (see id_severity/current_sha_for_id above) -
-            # one case arm per finding-producing loop in `compare`. Every
-            # pattern is anchored to the fixed label text `compare` itself
-            # emits at the START of body, never matched as a substring
-            # anywhere in it: a workspace-controlled value (a git config
-            # value, a path) can contain text like "workspace file: " or
-            # "TCB (sandbox self) modified: " and reach $body unmodified, so
-            # an unanchored match could be hijacked into deriving the wrong
-            # (attacker-chosen) id for the wrong namespace - approving it
-            # would then baseline an id that can never match, un-baselining
-            # the real finding while still counting it as acknowledged. A
-            # finding whose body matches none of these (currently only
-            # CROSS-PROJECT) has no id to record, so it's shown but not
-            # offered a/s.
+            # Derive the namespaced id baseline_suppresses() compares
+            # against - one arm per finding-producing loop in `compare`.
             #
-            # The id itself must then be recovered with a shortest-prefix
-            # strip of that same fixed label (`${body#LABEL}`), never
-            # `${body##*: }`: the latter strips up to the LAST ": " in the
-            # whole string, and a workspace-controlled path can itself
-            # contain a literal ": " (e.g. ".git/hooks/pre-commit: decoy"),
-            # which would truncate the derived id to everything after that
-            # planted separator - the case match being anchored doesn't
-            # protect the extraction that follows it.
+            # Both steps must resist workspace-controlled text in $body. The
+            # patterns anchor to the fixed label at the START of body: a git
+            # config value or path can contain "workspace file: ", so an
+            # unanchored match derives an attacker-chosen id in the wrong
+            # namespace, and approving it baselines an id that never matches
+            # while counting the real finding as acknowledged. The id is then
+            # recovered with `${body#LABEL}`, never `${body##*: }` - the
+            # latter strips to the LAST ": ", and a path can contain one.
+            # A body matching no arm (currently only CROSS-PROJECT) gets no
+            # id: shown, but not offered a/s.
             #
-            # Two forms come out of this, because the two consumers need
-            # different ones (see enc_id): $id is the ENCODED id exactly as
-            # the report carries it, and is what gets written to the
-            # baseline file, which stores that form. $rid is the same id
-            # decoded back to the raw bytes on disk, for current_sha_for_id
-            # to stat and hash and for id_severity to classify. Only the
-            # id-bearing portion is decoded - never the namespace prefix
-            # this code adds itself.
-            #
-            # $lbl and $surface are for DISPLAY only: the fixed label this
-            # arm matched, and the workspace-named thing inside the body that
-            # the finding is about. Splitting them is what lets the surface
-            # be coloured without colouring the audit's own words around it,
-            # and they are recomputed here rather than derived from $id
-            # because a gitconfig body carries `key=value` while its id is
-            # the key alone.
+            # $id is the ENCODED form the report and baseline file carry;
+            # $rid is it decoded, for current_sha_for_id to stat and hash and
+            # id_severity to classify. Only the id-bearing portion decodes,
+            # never the namespace prefix added here. $lbl and $surface are
+            # display-only, split so the surface can be coloured without the
+            # audit's own words around it, and recomputed rather than derived
+            # from $id because a gitconfig body carries `key=value` while its
+            # id is the key alone.
             ackable=0; id=""; rid=""; lbl=""; surface=""
+            # Host-execution surfaces first, because their labels come from
+            # the harness plugin rather than from a literal here. Matched by
+            # an exact prefix strip - "$body" with the label removed differs
+            # from "$body" only when the label really is at position 0 -
+            # rather than by a `case` pattern, which would treat a label's
+            # own characters as globs. Anchoring is the whole point (#45):
+            # an unanchored match lets chosen text derive an id in the wrong
+            # namespace, and approving it baselines an id that never matches
+            # while the real finding counts as acknowledged.
+            for _rec in ${DEJIMA_HOST_EXEC[@]+"${DEJIMA_HOST_EXEC[@]}"}; do
+                IFS=$'\t' read -r _ns _leaf _path _label <<< "$_rec"
+                _rest="${body#"$_label"}"
+                [ "$_rest" != "$body" ] || continue
+                ackable=1; lbl="$_label"; surface="$_rest"
+                id="$_ns:$_rest"; rid="$_ns:$(dec_id "$_rest")"
+                break
+            done
+            # Only if none of them claimed it: a host-exec label that
+            # matched has already derived the id, and the arms below must
+            # not get a second go at the same text.
+            if [ -z "$id" ]; then
             case "$body" in
                 "HOST-CODE-EXECUTION workspace file: "*)
                     ackable=1; lbl="HOST-CODE-EXECUTION workspace file: "
@@ -1509,18 +1652,14 @@ case "$cmd" in
                     cfgline="${body#changed git config: }"
                     ekey="${cfgline%%=*}"; surface="$ekey"
                     id="gitconfig:$ekey"; rid="gitconfig:$(dec_id "$ekey")" ;;
-                "HOST-CODE-EXECUTION statusline.sh changed (runs UNSANDBOXED on the host via settings.json statusLine.command): "*)
-                    ackable=1; lbl="HOST-CODE-EXECUTION statusline.sh changed (runs UNSANDBOXED on the host via settings.json statusLine.command): "
-                    esl="${body#HOST-CODE-EXECUTION statusline.sh changed (runs UNSANDBOXED on the host via settings.json statusLine.command): }"
-                    surface="$esl"
-                    id="statusline:$esl"; rid="statusline:$(dec_id "$esl")" ;;
-                "TCB (sandbox self) modified: "*)
-                    ackable=1; lbl="TCB (sandbox self) modified: "
-                    etcb="${body#TCB (sandbox self) modified: }"
+                "TCB (sandbox self) modified, INVISIBLE to git: "*)
+                    ackable=1; lbl="TCB (sandbox self) modified, INVISIBLE to git: "
+                    etcb="${body#TCB (sandbox self) modified, INVISIBLE to git: }"
                     surface="$etcb"
                     id="tcb:$etcb"; rid="tcb:$(dec_id "$etcb")" ;;
                 *) ackable=0 ;;
             esac
+            fi
             # Defence in depth: even an anchored arm above should never
             # leave id empty while claiming ackable, but don't trust that -
             # crit_acked (and thus whether a report can shed -CRITICAL.txt)
@@ -1545,6 +1684,18 @@ case "$cmd" in
                 if [ "$tier" = CRITICAL ]; then crit_acked=$((crit_acked + 1)); fi
                 continue
             fi
+            # A finding cleared on an earlier pass because its surface was
+            # gone. Nothing was recorded for it - there is no content to pin -
+            # so baseline_suppresses cannot cover it, and it is ackable, so
+            # the un-ackable branch below does not either. Without this it
+            # re-presents on every run: the same forever-loop the clear was
+            # meant to end, one step further along.
+            if [ "$ackable" -eq 1 ] && [ "$report_reviewed" -eq 1 ] \
+               && [ -z "$(current_sha_for_id "$rid")" ]; then
+                already=$((already + 1))
+                if [ "$tier" = CRITICAL ]; then crit_acked=$((crit_acked + 1)); fi
+                continue
+            fi
             # An un-ackable finding in a report already walked to the end.
             # There is nothing to record for one, so without this it would
             # re-prompt on every run forever, gating the exit from `ack` on a
@@ -1611,18 +1762,25 @@ case "$cmd" in
                     else
                         sha="$(current_sha_for_id "$rid")"
                         if [ -z "$sha" ]; then
-                            # Nothing left to pin the approval to: current_sha_for_id
-                            # returning empty means $id no longer exists (deleted
-                            # between compare and ack, or - for gitconfig - the key
-                            # was unset). Writing `safe <id> ` anyway would record a
-                            # baseline entry baseline_suppresses() can never match
-                            # (its own empty-sha guard rejects it as malformed), so
-                            # the surface would stay permanently unsuppressible -
-                            # while still advancing crit_acked and letting a CRITICAL
-                            # report shed its -CRITICAL.txt protection. Refuse
-                            # instead: record nothing, and leave the report's
-                            # protection exactly as it was.
-                            echo "  cannot approve: $id no longer exists"
+                            # The surface is gone - deleted since compare, or a
+                            # gitconfig key unset. There is nothing to pin an
+                            # approval to, and `safe <id> ` with an empty sha would
+                            # record a baseline line baseline_suppresses() can never
+                            # match, leaving the surface permanently unsuppressible.
+                            #
+                            # So nothing is recorded - but the finding is CLEARED,
+                            # because a surface that does not exist does not run.
+                            # Refusing outright is what made one deleted file pin
+                            # the audit forever: crit_acked never reached crit_seen,
+                            # the report kept -CRITICAL.txt, and select_report picks
+                            # CRITICAL first, so every other report stayed hidden
+                            # behind a finding that could not be resolved by any
+                            # means. If the surface comes back, it comes back as a
+                            # new finding, which is the honest outcome.
+                            echo "  ${C_OK}cleared${C_RESET}: $(sanitize "$id") no longer exists"
+                            removed_cleared=1
+                            [ "$tier" = CRITICAL ] && crit_acked=$((crit_acked + 1))
+                            sc_done+=("$rid")
                         else
                             pending+=("safe $id $sha")
                             sc_done+=("$rid")
@@ -1643,8 +1801,9 @@ case "$cmd" in
         # Written once the report has been walked to its end, and only when
         # it actually showed something that cannot be acknowledged - there is
         # nothing else to stop those from re-prompting forever.
-        if [ "$unackable_shown" -eq 1 ] && [ "$report_reviewed" -eq 0 ]; then
-            pending+=("reviewed $(basename "$report")")
+        if { [ "$unackable_shown" -eq 1 ] || [ "$removed_cleared" -eq 1 ]; } \
+           && [ "$report_reviewed" -eq 0 ]; then
+            pending+=("reviewed $(report_key "$report")")
         fi
         if [ "${#pending[@]}" -gt 0 ]; then
             mkdir -p "$BASELINE_DIR"
@@ -1682,10 +1841,33 @@ case "$cmd" in
         else
             echo "dejima audit: nothing recorded."
         fi
+
+        # Next report, if there is one. Asked rather than assumed: an operator
+        # who has decided enough for now should not have to abort out of a
+        # report they have not read to stop.
+        next_report="$(select_report "$report_dir")" || next_report=""
+        if [ -n "$next_report" ] && [ "$next_report" != "$report" ]; then
+            read -r n_all n_crit <<< "$(count_reports "$report_dir")"
+            printf '\n%s report(s) still waiting' "$n_all"
+            [ "$n_crit" -gt 0 ] && printf ', %s%s with unacknowledged CRITICAL%s' \
+                "$C_CRIT" "$n_crit" "$C_RESET"
+            printf '. Continue? [%sy%s/%sn%s] ' "$C_KEY" "$C_RESET" "$C_KEY" "$C_RESET"
+            # fd 3, the same duplicate of the original stdin the finding
+            # prompt reads: the report is this loop's stdin, and /dev/tty is
+            # not the caller's input when `ack` is driven by a pipe.
+            { read -r answer <&3; } 2>/dev/null || answer=n
+            printf '\n'
+            case "$answer" in
+                y|Y|"") report="$next_report"; sc="$(sidecar_dir "$report")"; continue ;;
+            esac
+        fi
+        break
+        done
         exit 0
         ;;
     *)
         echo "usage: dejima-audit {snapshot|compare|show|ack} [workspace] [state-file]" >&2
+        echo "  show [--full]   index of waiting reports; --full prints one" >&2
         echo "  workspace defaults to the enclosing git repository, else the current directory" >&2
         exit 2
         ;;

bin/dejima-config-lib.sh

@@ -22,33 +22,27 @@
 # Schema
 #
 # Three TAB-separated columns: key, type, default. One source of truth - both
-# _config_spec and _config_keys read this, so a key cannot exist for lookup
-# but be missing from enumeration (which is what --print-config walks).
+# _config_spec and _config_keys read it, so a key cannot be lookup-able but
+# missing from --print-config's enumeration.
 #
-# A "?" prefix on the type means the empty string is a legal value. Without
-# it, empty is rejected: the distinction is between a setting that is
-# meaningfully absent (ssh_key, loki_url) and one that must always have a
+# A "?" type prefix means empty is legal, distinguishing a setting that is
+# meaningfully absent (ssh_key, loki_url) from one that must always have a
 # value (firewall, share_dir).
 #
-# The three axis keys are `provider:<axis>` - validated by looking the name
-# up under the provider root (see _config_extra_type_valid below and
-# dejima-provider-lib.sh) rather than a fixed enum, so adding a provider is
-# adding a directory, not editing this schema.
-#
-# [forge.<name>] and [harness.<name>] keys are absent from this table for the
-# same reason: enumerating a forge's own keys here would mean a new forge
-# needs a core edit, which is exactly what the provider axes are for.
-# Those two sections are resolved instead through the _config_extra_*
-# functions below, stubbed here and overridden by dejima-provider-lib.sh once
-# it is sourced, against the type+default each provider's own `keys` manifest
-# entry declares (see _provider_keyspec_parse there).
+# The axis keys are `provider:<axis>`, validated against the provider root
+# rather than a fixed enum, so adding a provider is adding a directory.
+# [forge.<name>] and [harness.<name>] keys are absent here for the same
+# reason - enumerating them would make a new forge need a core edit. They
+# resolve through the _config_extra_* functions below, stubbed here and
+# overridden by dejima-provider-lib.sh, against each provider's own `keys`
+# manifest entry.
 # ---------------------------------------------------------------------------
 _config_table() {
     cat <<'EOF'
-core.harness	provider:harness	claude-code
+core.harness	provider:harness
 core.forge	provider:forge	none
 core.workflow	provider:workflow	local-only
-core.command	identifier	claude
+core.command	?identifier
 core.firewall	bool	on
 core.share_dir	path	~/.local/share/dejima/shared
 core.ssh_key	?path
@@ -61,6 +55,9 @@ credentials.proxy	bool	on
 credentials.transport	enum:tls|loopback	tls
 logging.loki_url	?url
 logging.conversation	bool	off
+epitaph.enabled	bool	on
+epitaph.dir	path	~/.local/share/dejima/epitaphs
+epitaph.loki	bool	off
 audit.dir	path	~/.local/share/dejima/audit
 audit.baselines_dir	path	~/.local/share/dejima/baselines
 audit.keep	int:1:1000	10
@@ -313,7 +310,7 @@ _config_extra_section_known() { return 1; }
 
 _config_section_known() {
     case "$1" in
-        core|github|mounts|credentials|logging|audit|remote-access) return 0 ;;
+        core|github|mounts|credentials|logging|audit|remote-access|epitaph) return 0 ;;
         *) _config_extra_section_known "$1" ;;
     esac
 }

bin/dejima-epitaph

@@ -0,0 +1,171 @@
+#!/bin/bash
+# Epitaph: what a session leaves behind when its context dies.
+#
+# Claude Code writes a summary each time a session is compacted. It is the
+# only durable account of what a long session worked out, and it goes with
+# the container. This keeps it.
+#
+# Runs entirely on the host, from bin/dejima's exit trap. Nothing in the
+# container writes here and the store is never mounted, so a session cannot
+# read what earlier sessions concluded or edit what this one is about to
+# record - the same arrangement as the audit store, and for the same reason
+# (#104: anything a session writes and a later one reads is a channel into
+# that later session's premises).
+#
+# The transcript IS session-writable, so a summary is only as honest as the
+# file it came from. This preserves what was recorded, it does not attest to
+# it.
+#
+# Usage: dejima-epitaph save <workspace> [transcripts-dir]
+set -uo pipefail
+
+# shellcheck source=dejima-provider-lib.sh
+source "$(dirname "$(readlink -f "$0")")/dejima-provider-lib.sh"
+# shellcheck source=dejima-slug-lib.sh
+source "$(dirname "$(readlink -f "$0")")/dejima-slug-lib.sh"
+config_validate || exit 1
+
+cmd="${1:-}"; ws_arg="${2:-}"; transcripts="${3:-}"
+
+[ "$(config_get epitaph enabled)" = "on" ] || exit 0
+STORE="$(config_get epitaph dir)"
+[ -n "$transcripts" ] || transcripts="$(harness_config_get projects_dir)"
+
+ws="$ws_arg"
+[ -n "$ws" ] || ws="$(pwd)"
+ws="$(cd "$ws" 2>/dev/null && pwd)" || exit 0
+
+case "$cmd" in
+    save) ;;
+    *) echo "usage: dejima-epitaph save <workspace> [transcripts-dir]" >&2; exit 2 ;;
+esac
+
+# One file per summary, named for the moment it was written, so a directory
+# listing is a timeline. Keyed on the summary record's own uuid: `save` runs
+# on every exit, and a session compacted three times has three epitaphs, two
+# of which an earlier run already wrote.
+mkdir -p "$STORE" 2>/dev/null || exit 0
+chmod 700 "$STORE" 2>/dev/null || true
+
+saved=0
+while IFS= read -r slug; do
+    [ -n "$slug" ] || continue
+    dir="$transcripts/$slug"
+    [ -d "$dir" ] || continue
+    out="$STORE/$slug"
+    for f in "$dir"/*.jsonl; do
+        [ -f "$f" ] || continue
+        # python3 is already a hard dependency (install.sh refuses without
+        # it), and this is line-delimited JSON with embedded newlines in the
+        # payload - not something to take apart with sed.
+        n="$(python3 - "$f" "$out" <<'PY'
+import json, os, sys
+
+path, out = sys.argv[1], sys.argv[2]
+pending = None   # metadata from the compact_boundary that precedes a summary
+written = 0
+
+for line in open(path, errors="replace"):
+    line = line.strip()
+    if not line or not line.startswith("{"):
+        continue
+    try:
+        rec = json.loads(line)
+    except ValueError:
+        continue
+
+    if rec.get("subtype") == "compact_boundary":
+        pending = rec.get("compactMetadata") or {}
+        continue
+
+    if not rec.get("isCompactSummary"):
+        continue
+
+    body = rec.get("message", {}).get("content")
+    if isinstance(body, list):
+        body = "\n".join(p.get("text", "") for p in body if isinstance(p, dict))
+    if not isinstance(body, str) or not body.strip():
+        pending = None
+        continue
+
+    uid = rec.get("uuid") or ""
+    stamp = (rec.get("timestamp") or "").replace(":", "").replace("-", "")[:15]
+    name = "%s-%s.md" % (stamp or "unknown", uid[:8] or "nouuid")
+    dest = os.path.join(out, name)
+    if os.path.exists(dest):
+        pending = None
+        continue
+
+    meta = pending or {}
+    head = [
+        "# Epitaph",
+        "",
+        "- session: %s" % (rec.get("sessionId") or "unknown"),
+        "- at: %s" % (rec.get("timestamp") or "unknown"),
+        "- trigger: %s" % (meta.get("trigger") or "unknown"),
+    ]
+    if meta.get("preTokens") is not None:
+        head.append("- tokens: %s before, %s after"
+                    % (meta.get("preTokens"), meta.get("postTokens")))
+    head += [
+        "",
+        "Written by the session and preserved verbatim. The transcript it came",
+        "from is writable from inside the container, so this records what was",
+        "there rather than vouching for it.",
+        "", "---", "",
+    ]
+    tmp = dest + ".part"
+    os.makedirs(out, exist_ok=True)
+    with open(tmp, "w") as fh:
+        fh.write("\n".join(head) + body.rstrip() + "\n")
+    os.chmod(tmp, 0o600)
+    os.rename(tmp, dest)
+    written += 1
+    pending = None
+
+print(written)
+PY
+)" || n=0
+        saved=$((saved + ${n:-0}))
+    done
+    # The directory is made by the first write into it, not by looking at the
+    # slug: most worktree slugs have transcripts and no compaction summary in
+    # any of them, so creating one per slug up front filled the store with
+    # empty directories and made it look like it held something for each. The
+    # rmdir clears the ones earlier runs already left; it cannot touch a slug
+    # that kept anything.
+    rmdir "$out" 2>/dev/null || true
+done < <(workspace_slugs_sorted "$ws")
+
+[ "$saved" -gt 0 ] || exit 0
+echo "dejima: wrote $saved epitaph(s) to $STORE" >&2
+
+# Optional, and off by default: the store is the durable copy, this is for
+# operators who already centralise session logs. Host-side rather than
+# through the container's log-shipper, so the same "never in the session's
+# reach" property holds for the shipping path too.
+[ "$(config_get epitaph loki)" = "on" ] || exit 0
+loki="$(config_get logging loki_url)"
+[ -n "$loki" ] || exit 0
+while IFS= read -r slug; do
+    [ -n "$slug" ] || continue
+    for f in "$STORE/$slug"/*.md; do
+        [ -f "$f" ] && [ ! -f "$f.shipped" ] || continue
+        if python3 - "$loki" "$slug" "$f" <<'PY'
+import json, sys, time, urllib.request
+
+loki, slug, path = sys.argv[1], sys.argv[2], sys.argv[3]
+body = json.dumps({"streams": [{
+    "stream": {"job": "dejima-epitaph", "project": slug},
+    "values": [[str(time.time_ns()), open(path, errors="replace").read()]],
+}]}).encode()
+req = urllib.request.Request(loki.rstrip("/") + "/loki/api/v1/push", data=body,
+                             headers={"Content-Type": "application/json"})
+urllib.request.urlopen(req, timeout=10).read()
+PY
+        then
+            : > "$f.shipped"
+        fi
+    done
+done < <(workspace_slugs_sorted "$ws")
+exit 0

bin/dejima-harness-lib.sh

@@ -0,0 +1,241 @@
+# Shared library: loading a harness plugin, and the interface it declares
+# through. Sourced by bin/dejima. Not on PATH and not executable.
+#
+# Three tiers, and which one a thing belongs in is the whole design:
+#
+#   core      - the mechanism, and every enforcement point.
+#   manifest  - data. harnesses/<name>/harness.conf. Never sourced, never
+#               executed. Selects among named core mechanisms, supplies paths.
+#   plugin    - bin/harness/<name>.sh. The harness-specific remainder, as
+#               code, in this repository, supported one harness at a time.
+#
+# A feature belongs in the manifest if it can be stated as a value. It moves
+# here only when it cannot: opencode resolving its state directory through a
+# database, Claude Code pairing file-history to transcripts by session id.
+# "Add a manifest key" is the answer until it demonstrably is not.
+#
+# Plugins live under bin/ rather than beside their manifest because
+# dejima-audit's tcb_snapshot covers install.sh, bin/, hooks/, docker/ and
+# test/ - and not harnesses/. Host-side plugin code under harnesses/ would be
+# code that runs on the host and that a session could edit in dejima's own
+# checkout without raising a CRITICAL, which is the exact class the audit
+# exists to catch. Under bin/ it is covered for free, and "manifests are
+# data, never executed" (docs/security.md) stays true word for word.
+#
+# Plugins are SOURCED, not invoked. They are part of the trusted computing
+# base, so the risk worth designing against is not a careless plugin - it is
+# three plugins implementing the security model three different ways.
+# Sourcing lets a plugin use core's own primitives (_refuse_guarded_mount,
+# sanitize, ws_slug, harness_config_get) rather than re-deriving them, which
+# is what keeps one security model instead of one per harness. An invoked
+# protocol would also put a new text parser on the launch path, and every
+# parsing bug this repository has had is in that family.
+#
+# The rule that replaces isolation: A PLUGIN COMPUTES, CORE ENFORCES.
+# A plugin declares through the emit_* functions below, which validate at the
+# point of declaration. It never touches extra_mounts itself - a static test
+# holds that, because a plugin that appended directly would skip the guarded-
+# store refusal, and would skip it differently in each harness.
+
+# The hooks core will call. A plugin defines the ones its harness needs and
+# omits the rest; anything else it defines is a name core does not call, and
+# the static test refuses it rather than letting it look wired up.
+DEJIMA_HARNESS_HOOKS="harness_state_mounts harness_register_notify harness_register_guards harness_host_exec_surfaces"
+
+# What an operator loses when a harness cannot host a given guard. The set
+# itself is DEJIMA_KNOWN_GUARDS in dejima-provider-lib.sh, next to the
+# manifest schema and the validation that rejects a name outside it: a
+# constant the validation depends on must live where the validation does, or
+# a caller that loads only provider-lib validates against an empty set and
+# rejects every guard there is.
+_guard_description() {
+    case "$1" in
+        bash-deny)   printf 'obviously destructive shell commands are not refused' ;;
+        push-guard)  printf 'pushes to hosts and branches outside the allowed set are not refused' ;;
+        merge-guard) printf 'the agent can merge its own pull requests' ;;
+        notify)      printf 'remote access gets no waiting/done notifications' ;;
+        *)           printf 'unknown guard' ;;
+    esac
+}
+
+# Whether the selected harness declared a given guard. Core gates the work
+# on this rather than on whether the plugin happens to define a hook: a
+# harness that forgot the declaration but implemented the hook would get a
+# guard the launch message has just told the operator it does not have.
+harness_has_guard() {
+    case " $(manifest_get harness "$DEJIMA_HARNESS_NAME" guards) " in
+        *" $1 "*) return 0 ;;
+        *) return 1 ;;
+    esac
+}
+
+# Says which guards the selected harness cannot host. Silent when it hosts
+# them all, which is the common case and must not become noise an operator
+# learns to scroll past - the point is that the line only appears when
+# something really is missing.
+harness_report_guards() {
+    local harness="$1" declared guard missing=()
+    declared="$(manifest_get harness "$harness" guards)"
+    for guard in $DEJIMA_KNOWN_GUARDS; do
+        case " $declared " in
+            *" $guard "*) ;;
+            *) missing+=("$guard") ;;
+        esac
+    done
+    [ "${#missing[@]}" -gt 0 ] || return 0
+    printf '  %sharness %s does not support %s of dejima'"'"'s guards:%s\n' \
+        "$C_WARN" "$harness" "${#missing[@]}" "$C_RESET" >&2
+    for guard in "${missing[@]}"; do
+        printf '    %s%s%s - %s\n' "$C_KEY" "$guard" "$C_RESET" "$(_guard_description "$guard")" >&2
+    done
+    printf '  The container still bounds this session; these are the fast-fail\n' >&2
+    printf '  guards it does not get. See docs/development.md.\n' >&2
+}
+
+# Sources the selected harness's plugin. Hard error when it is missing:
+# support is per harness and deliberate (see "Adding a new harness" in
+# docs/development.md), so a harness with a manifest and no plugin is a
+# half-finished integration rather than a lightweight one.
+harness_plugin_load() {
+    local name="$1" plugin
+    [ -n "$name" ] || return 0
+    # Resolved from this file's own location, not from a caller's SCRIPT_DIR:
+    # install.sh sets that to the repo root while bin/dejima sets it to bin/,
+    # and a library that only works for one of its callers is a trap for the
+    # next one.
+    plugin="$(cd "${BASH_SOURCE[0]%/*}" && pwd)/harness/$name.sh"
+    if [ ! -f "$plugin" ]; then
+        echo "No harness plugin for '$name' at $plugin." >&2
+        echo "A supported harness has both a manifest and a plugin; see" >&2
+        echo "\"Adding a new harness\" in docs/development.md." >&2
+        exit 1
+    fi
+    # shellcheck disable=SC1090
+    source "$plugin"
+}
+
+# Calls one hook of the loaded plugin, if it defines it. Absence is normal:
+# not every harness needs every hook, and the manifest is where a harness
+# says what it cannot do at all.
+harness_hook() {
+    local hook="$1"; shift
+    declare -F "$hook" >/dev/null 2>&1 || return 0
+    "$hook" "$@"
+}
+
+# --- what a plugin declares through --------------------------------------
+#
+# Each of these validates and then appends to core's own state. A plugin
+# calls them; it does not reach past them.
+
+# A host path mounted into the container at the same path. $2 is the mode
+# ("ro", or empty for read-write).
+#
+# The guarded-store refusal runs here, at the point of declaration, so it
+# cannot be skipped by a plugin that forgot it: a harness state directory
+# that overlapped the approval or audit or epitaph store would otherwise
+# hand a session exactly what those stores exist to withhold, and a harness
+# is precisely the kind of thing whose paths are computed rather than
+# written down.
+emit_mount() {
+    local path="$1" mode="${2:-}"
+    [ -n "$path" ] || return 0
+    case "$path" in
+        *:*)
+            echo "Harness plugin emitted a mount path containing a colon: $path" >&2
+            exit 1 ;;
+    esac
+    _guarded_stores
+    _refuse_guarded_mount "harness state mount" "$path" \
+        ${GUARDED_STORES[@]+"${GUARDED_STORES[@]}"}
+    if [ -n "$mode" ]; then
+        extra_mounts+=(-v "$path:$path:$mode")
+    else
+        extra_mounts+=(-v "$path:$path")
+    fi
+}
+
+# An environment variable for the container. Name-checked because it becomes
+# a `-e NAME=value` argument, and a name carrying a space or an `=` would
+# split into something else entirely.
+emit_env() {
+    local name="$1" value="${2:-}"
+    case "$name" in
+        ''|*[!A-Za-z0-9_]*)
+            echo "Harness plugin emitted an invalid environment variable name: $name" >&2
+            exit 1 ;;
+    esac
+    extra_mounts+=(-e "$name=$value")
+}
+
+# --- host-execution surfaces ---------------------------------------------
+#
+# Paths a harness causes to run ON THE HOST, unsandboxed, from configuration
+# a session can reach. Claude Code's statusline is the instance that exists:
+# settings.json's statusLine.command names a script the host runs, so a
+# session that can write that file has a way onto the host that no container
+# boundary sees.
+#
+# Declared by the plugin rather than by the manifest, and the reason is a
+# trust boundary rather than convenience. dejima-audit's `ack` derives a
+# finding's id by stripping a FIXED label from the front of the finding text
+# - anchored deliberately, because an unanchored match lets chosen text
+# derive an id in the wrong namespace and baseline an approval that never
+# matches while the real finding is counted as acknowledged (#45). Those
+# labels are therefore part of a parser's grammar. Manifests are data, and
+# provider directories are NOT in tcb_snapshot's list, so in the self-mount
+# case a session can edit one without raising a CRITICAL. A plugin is under
+# bin/, which the audit does watch. The label stays a literal in trusted
+# code; it just lives with the harness that owns it.
+#
+# Each surface carries its own namespace, so ids stay stable per harness and
+# no rename is forced on anyone's baseline.
+DEJIMA_HOST_EXEC=()
+
+# namespace, leaf, path, label. The id is "<namespace>:<leaf>", which is what
+# a baseline stores, so neither may change for a harness already in use.
+emit_host_exec() {
+    local ns="$1" leaf="$2" path="$3" label="$4"
+    [ -n "$path" ] || return 0
+    case "$ns" in
+        ''|*[!a-z0-9-]*)
+            echo "Harness plugin emitted an invalid host-exec namespace: $ns" >&2
+            exit 1 ;;
+    esac
+    # The leaf lands in an id, which is compared against baseline entries and
+    # written into a report; a tab or newline in either would split a record.
+    case "$leaf$label" in
+        *$'\t'*|*$'\n'*)
+            echo "Harness plugin emitted a host-exec leaf or label containing a tab or newline" >&2
+            exit 1 ;;
+    esac
+    [ -n "$leaf" ] && [ -n "$label" ] || {
+        echo "Harness plugin emitted a host-exec surface with no leaf or label" >&2
+        exit 1
+    }
+    DEJIMA_HOST_EXEC+=("$ns"$'\t'"$leaf"$'\t'"$path"$'\t'"$label")
+}
+
+# Asks the loaded plugin for its surfaces. Empty when no harness is selected
+# or the harness has none, which is a normal state rather than an error.
+host_exec_load() {
+    DEJIMA_HOST_EXEC=()
+    harness_hook harness_host_exec_surfaces
+}
+
+# The host path behind an id, or nothing if that id is not a host-exec one.
+host_exec_path_for_id() {
+    local id="$1" rec ns leaf path label
+    for rec in ${DEJIMA_HOST_EXEC[@]+"${DEJIMA_HOST_EXEC[@]}"}; do
+        IFS=$'\t' read -r ns leaf path label <<< "$rec"
+        [ "$id" = "$ns:$leaf" ] || continue
+        printf '%s' "$path"
+        return 0
+    done
+    return 1
+}
+
+# Whether an id names a host-execution surface. Used where severity is
+# decided: every one of these is CRITICAL, because the whole category is
+# "runs on the host, outside everything".
+host_exec_is_id() { host_exec_path_for_id "$1" >/dev/null; }

bin/dejima-notify-watch

@@ -7,30 +7,24 @@
 #   (no argument)  watch until killed - what the systemd unit runs
 #   --once         one scan, then exit
 #
-# This script is the trust boundary. Everything under the marker root is
-# written by a sandboxed session that may be compromised, and the device it
-# ultimately reaches - the operator's phone - is outside the sandbox. The
-# in-container hook's own enum constrains nothing here: a session owns the
-# contents of its bind-mounted marker directory and can create any name it
-# likes in it, hook or no hook. The guarantee rests entirely on this file:
+# This script is the trust boundary: a possibly-compromised session writes
+# the markers, and the phone they reach is outside the sandbox. The
+# in-container hook's enum constrains nothing here - a session owns its
+# marker directory and can create any name it likes. The guarantee is here:
 #
-#   - the marker NAME is the only thing consulted, and only by iterating a
-#     literal enum, so a name outside it is never stat'd, read, or removed
-#   - marker CONTENT is never read, so there is nothing to inject
-#   - names are never interpolated into a command
-#   - `rm -f` unlinks a symlink instead of following it, so a marker aimed at
-#     a host file removes the link and leaves the file alone
+#   - only the marker NAME is consulted, by iterating a literal enum, so a
+#     name outside it is never stat'd, read, or removed
+#   - marker CONTENT is never read, and names never reach a command
+#   - `rm -f` unlinks a symlink rather than following it
 #   - the click URL and every ntfy header are composed from host-side data
 #
-# That last point is what the enum is really for. ntfy's own headers include
-# X-Actions (which can make the phone issue an arbitrary HTTP request or fire
-# an Android intent), X-Attach and X-Icon (which fetch URLs), X-Email and
-# X-Call. A single attacker-controlled header here would turn the phone into
-# a confused deputy inside the operator's network.
+# The last point is what the enum is for: ntfy's X-Actions can make the phone
+# issue an HTTP request or fire an Android intent, X-Attach and X-Icon fetch
+# URLs. One attacker-controlled header makes the phone a confused deputy
+# inside the operator's network.
 #
-# One scan covers every run directory, so nothing a single session can do to
-# its own markers may abort it: that would let one compromised sandbox silence
-# every other session's notifications.
+# One scan covers every run directory, so nothing a session does to its own
+# markers can abort it and silence every other session's notifications.
 set -euo pipefail
 
 SCRIPT_DIR="$(cd "$(dirname "$(readlink -f "${BASH_SOURCE[0]}")")" && pwd)"

bin/dejima-project-lib.sh

@@ -115,24 +115,20 @@ read_declared() {
 # Shared argument parsing for dejima-allow and dejima-dep, so the two CLIs
 # cannot drift on what --project means.
 #
-# The hard part is that --project's path is OPTIONAL, which makes
-# `--project X` ambiguous on its own: X could be the path, or the required
-# positional. Deciding by "is X the last token" is wrong for `dejima-dep list
-# --project ~/p`, which has no trailing positional to protect and would
-# silently resolve to the current directory instead - the scope quietly
-# becoming one the caller did not name. So the decision is made against how
-# many positionals the command actually requires after the flag:
+# --project's path is OPTIONAL, so `--project X` is ambiguous: X could be the
+# path or the required positional. Deciding by "is X the last token" breaks
+# `dejima-dep list --project ~/p`, which has no trailing positional and would
+# silently scope to the current directory instead. So:
 #
-#   --project takes the next token as its path only if doing so still leaves
-#   the command its required trailing positional(s).
+#   --project takes the next token as its path only if that still leaves the
+#   command its required trailing positional(s).
 #
-# Required counts are per-command constants declared by the caller, not
-# inferred here.
+# Required counts are per-command constants the caller declares.
 #
-# Sets dj_global, dj_project, dj_project_set, dj_container and the
-# dj_positional array. dj_project_set exists because "--project with no path"
-# (the current directory) and "no --project at all" are different scopes and
-# an empty dj_project cannot tell them apart.
+# Sets dj_global, dj_project, dj_project_set, dj_container and dj_positional.
+# dj_project_set exists because "--project with no path" (the current
+# directory) and "no --project at all" are different scopes that an empty
+# dj_project cannot distinguish.
 parse_scope_args() {
     local required="$1"; shift
     local -a args=("$@")
@@ -336,34 +332,26 @@ declare_entry() {
     local project="$1" kind="$2" entry="$3" validator
     validator="$(_approved_validator "$kind")" || return 1
     "$validator" "$entry" || { echo "Invalid $kind entry: $entry" >&2; return 1; }
-    # This is the one write in this file that lands on the session-writable
-    # bind mount, so it is the one that must not follow a symlink. A
-    # compromised session can plant one and then ask the user - through the
-    # startup briefing, which tells it to hand over exactly this command - to
-    # run the CLI that trips it. Following it would aim a host-privileged
-    # append anywhere the user can write: at ~/.config/dejima/allowed-domains,
-    # silently turning a project approval into a global one; at
-    # ~/.config/dejima/dependencies, where the entry is installed as root at
-    # image build, on the one code path with no egress firewall in front of it;
-    # or at any host file at all.
+    # The one write here that lands on the session-writable bind mount, so
+    # the one that must not follow a symlink. A session can plant one and
+    # then ask the user - via the briefing, which hands over exactly this
+    # command - to run the CLI that trips it, aiming a host-privileged
+    # append at ~/.config/dejima/allowed-domains (project approval becomes
+    # global) or dependencies (installed as root at image build, the one
+    # path with no firewall in front of it).
     #
-    # A test would not be enough. The briefing tells the agent to have the user
-    # run `dejima-allow --project` from a terminal outside the sandbox, so the
-    # session is normally still LIVE while this runs - and since it is the one
-    # asking for the command, it picks the moment and can spin a rename loop
-    # across any window between a check and the write. So there is no check:
-    # the guard is the open itself. O_NOFOLLOW makes a symlink at either path
-    # fail the open outright, the directory is opened O_DIRECTORY and the file
-    # is opened relative to that fd, so neither can be swapped for the other's
-    # sake afterwards, and the whole read-then-append happens through that one
-    # fd. A session that wins the swap gets an open of the object it replaced,
-    # not a redirected write.
+    # Checking first is not enough: the session is normally still live when
+    # the user runs this, and it chose the moment, so it can spin a rename
+    # loop across any check-to-write window. The guard is the open itself.
+    # O_NOFOLLOW fails outright on a symlink, the directory is opened
+    # O_DIRECTORY and the file relative to that fd, and the whole
+    # read-then-append goes through it - so winning the swap gets an open of
+    # the replaced object, not a redirected write.
     #
-    # bash cannot open a path that way, so this is python3 - a hard dependency
-    # already (install.sh refuses to install without it). O_NONBLOCK and the
-    # S_ISREG check cover the non-symlink shapes: a FIFO planted here would
-    # otherwise block a host CLI the user was told to run for as long as the
-    # session declines to open the other end.
+    # bash cannot open a path that way, hence python3 (already a hard
+    # dependency). O_NONBLOCK and the S_ISREG check cover the non-symlink
+    # shapes: a planted FIFO would otherwise hang the host CLI for as long as
+    # the session declined to open the other end.
     python3 - "$project/.dejima" "$kind" "$entry" "$DEJIMA_MAX_DECLARED" <<'PY'
 import errno, os, stat, sys
 

bin/dejima-provider-lib.sh

@@ -141,6 +141,10 @@ EOF
             ;;
         harness) cat <<'EOF'
 command	identifier	claude
+command_path	?path
+guards	?list:identifier
+no_confirm_args	?template
+settings_args	?template
 keys	?list:keyspec
 resume_command	?template
 briefing	?relpath
@@ -329,28 +333,17 @@ manifest_get() {
 # directory whose manifest declares its own keys, not editing core.
 # ---------------------------------------------------------------------------
 
-# config_validate re-parses the WHOLE config file on every config_get call
-# (see its own comment - "on the order of thirty reads" per launch), and each
-# pass through a [forge.<name>]/[harness.<name>] line calls back into here.
-# Going through manifest_get/manifest_validate here would mean thirty-odd
-# full-manifest validations (every field, not just `keys`) for every key in
-# those sections - so this reads and validates only the `keys` field itself,
-# a plain line-scan plus the one type check _provider_valid_value already
-# does for it, skipping fields (pr_command, capabilities, env_export, ...)
-# that this lookup never uses. The manifest's OTHER fields still get their
-# full manifest_validate elsewhere before anything consequential reads them
-# (provider_validate_selection, forge_hosts, forge_env_exports,
-# briefing_compose) - this function only ever feeds a config schema type
-# check, never a forge action.
+# config_validate re-parses the whole config on every config_get (~thirty per
+# launch), and each [forge.<name>] line calls back here. Going through
+# manifest_validate would mean that many full-manifest validations, so this
+# line-scans the `keys` field alone. Every other field still gets a full
+# manifest_validate before anything consequential reads it - this feeds a
+# config schema type check, never a forge action.
 #
-# Cached per axis+name for the rest of THIS command substitution: a
-# manifest cannot change mid-launch, so re-deriving it for every key in the
-# same section within one config_validate pass is pure waste. The cache
-# itself does not survive past this subshell (bash forks a new one per
-# config_get call), so it only dedupes the N keys of one section against
-# each other, not across a launch's ~thirty reads - a cheap, safe partial
-# win rather than a cross-process cache this file has no story for
-# invalidating.
+# Cached per axis+name for this command substitution only. bash forks a new
+# subshell per config_get, so it dedupes the keys of one section against each
+# other rather than across a launch: a cheap partial win, not a cross-process
+# cache this file has no way to invalidate.
 declare -gA _PROVIDER_KEYS_CACHE
 _provider_keys_cached() {
     local axis="$1" name="$2" cache_key="$axis:$name"
@@ -371,6 +364,10 @@ _provider_keys_cached() {
 # here is a per-line fork thirty times over.
 _provider_keys_field() {
     local axis="$1" name="$2" dir file line key value="" seen=0 type n=0
+    # An unselected axis contributes no keys - a different question from
+    # whether a selection is valid. Without this, --print-config prints
+    # "invalid harness name:" and fails at the thing it is for.
+    [ -n "$name" ] || return 1
     dir="$(provider_dir "$axis" "$name")" || return 1
     file="$dir/$(_provider_manifest_name "$axis")"
     [ -f "$file" ] || return 1
@@ -433,6 +430,37 @@ _config_extra_spec() {
     return 1
 }
 
+# One key of the SELECTED harness, rather than of a harness named literally
+# at the call site.
+#
+# `keys` is declared per harness, so reading one through a fixed section name
+# makes it a declaration only one harness can ever be read through: a second
+# harness gets selected, gets validated, and is then handed the first one's
+# paths with nothing erroring anywhere. That is a silent wrong answer rather
+# than a failure, which is why it is worth a helper rather than a longer
+# expression at each site.
+#
+# Empty when no harness is selected. bin/dejima refuses to start in that
+# state (there is no default, deliberately - see [core] harness), so it is
+# reachable only from the standalone tools, where a harness-specific path
+# that resolves to nothing is the honest answer for "no harness chosen".
+harness_config_get() {
+    local name
+    name="$(config_get core harness)"
+    [ -n "$name" ] || return 0
+    config_get "harness.$name" "$1"
+}
+
+# Every guard rail dejima knows how to apply. A harness declares which of
+# them it can host, in its manifest's `guards`; the ones it cannot are named
+# at launch (see harness_report_guards) rather than quietly absent.
+#
+# Here rather than in dejima-harness-lib.sh because provider_validate_selection
+# rejects a name outside this set, and a constant the validation depends on
+# has to live where the validation does - providers.bats sources only this
+# file, and against an empty set every guard is unknown.
+DEJIMA_KNOWN_GUARDS="bash-deny push-guard merge-guard notify"
+
 # Enumerates the keys of the SELECTED forge/harness (core.forge/core.harness,
 # falling back to their schema default when no config file exists yet to
 # read a selection from - e.g. an isolated call against a fresh checkout)
@@ -472,6 +500,25 @@ provider_validate_selection() {
     manifest_validate workflow "$workflow" || return 1
     manifest_validate harness "$harness" || return 1
 
+    # Every guard rail dejima knows how to apply. A harness declares which of
+    # them it can host; the ones it cannot are named at launch rather than
+    # quietly absent, because an operator who has them everywhere else would
+    # otherwise have to read a manifest to find out they are gone here.
+    #
+    # A name outside this set is a hard error rather than an unknown guard
+    # treated as absent: a typo would otherwise silently turn a guard off,
+    # which is the failure this whole declaration exists to prevent.
+    local guard
+    for guard in $(manifest_get harness "$harness" guards); do
+        case " $DEJIMA_KNOWN_GUARDS " in
+            *" $guard "*) ;;
+            *)
+                echo "harness '$harness' declares an unknown guard: $guard" >&2
+                echo "Known guards: $DEJIMA_KNOWN_GUARDS" >&2
+                return 1 ;;
+        esac
+    done
+
     caps="$(manifest_get forge "$forge" capabilities)"
     requires="$(manifest_get workflow "$workflow" requires)"
     for need in $requires; do
@@ -726,6 +773,14 @@ What this means:
   .dejima/allowed-domains, so it is shared with everyone working on it
   rather than living on one machine. Quote that --container value exactly as
   written here - it is this container's real ID, not something to infer.
+  And a domain can be an action rather than a source: a deploy hook, a
+  payment or production API acts on the world the moment it is called, with
+  no diff for anyone to review afterwards. Do not ask for one of those to
+  finish a task - say what you need and let the operator decide.
+- The project directory is mounted whole, so it can hold things that are not
+  code: a .env, a database dump, a key someone left in the tree. Say so if
+  you find one, and do not open, copy or send it anywhere unless the task
+  actually requires it.
 - Stay within the task you were given. Since there are no permission
   prompts here, be more conservative than usual about anything destructive,
   irreversible, or outside the project directory.
@@ -750,6 +805,33 @@ _briefing_core_worktree() {
   `.worktrees/` to `.git/info/exclude` (not `.gitignore` - this is
   sandbox-local, not something to commit) so it doesn't clutter `git
   status` for the user's real checkout.
+- A worktree isolates the working tree, NOT `.git/config`, which is shared
+  with the parent checkout. A bare `git config <key> <value>` run from
+  inside a worktree writes to the user's real repository config, not to
+  anything scoped to your branch. Use `git -C <path> config` against a
+  throwaway repo when setting config for a test or a repro. This matters
+  beyond a stray `user.name`: `core.hooksPath`, `credential.helper` and
+  `!`-aliases all execute on the host, and the same unscoped command
+  reaches them.
+BRIEFING_EOF
+}
+
+# Only when GitHub is a push destination this session. The default is off for
+# a reason the agent cannot infer from having the ability: docs/security.md,
+# "GitHub push". An agent that finds a real bug in a dependency and helpfully
+# opens a pull request on it has not made a mistake about permissions - it
+# has done something the operator was supposed to decide, and added to a queue
+# maintainers are already drowning in. Saying so here is cheaper than
+# discovering it afterwards, and the flag is exactly when it becomes possible.
+_briefing_core_github_push() {
+    cat <<'BRIEFING_EOF'
+- GitHub push is enabled for this session, for the repository you were
+  pointed at and nothing else. Do NOT push a branch, open a pull request, or
+  file an issue on any other project, however good the change would be. An
+  unsolicited pull request is a cost imposed on maintainers who did not ask
+  for it, not a contribution, and it arrives under the operator's name. If
+  you find a real bug somewhere else, say so in your reply and let them
+  decide what to do about it.
 BRIEFING_EOF
 }
 
@@ -1013,6 +1095,10 @@ briefing_compose() {
     text+=$'\n\n'
     text+="$workflow_frag"
     text+=$'\n\n'
+    if [ "${DEJIMA_GITHUB_PUSH:-off}" = "on" ]; then
+        text+="$(_briefing_core_github_push)"
+        text+=$'\n\n'
+    fi
     text+="$(_briefing_core_closing_a)"
     text+=$'\n\n'
     text+="$harness_frag"

bin/dejima-slug-lib.sh

@@ -0,0 +1,81 @@
+# Shared library: the ~/.claude/projects slug scheme, and which slugs belong
+# to a given workspace. Sourced by bin/dejima (to decide what to mount) and
+# bin/dejima-epitaph (to decide whose transcripts to keep). Not on PATH and
+# not executable.
+#
+# One definition, because the callers must not disagree. A slug the mount
+# exposes but a caller treats as foreign is a session whose work is silently
+# left behind; a slug a caller treats as the workspace's own but the mount
+# does not expose is a hole that nothing reports.
+
+# Claude Code's own project key: the absolute cwd with every non-alphanumeric
+# replaced by a dash. Matches docker/log-shipper.sh's Loki key and
+# dejima-project-lib.sh's project_slug, which cannot source this file (one
+# runs inside the container, the other is a separate library).
+ws_slug() { printf '%s' "$1" | sed 's/[^a-zA-Z0-9]/-/g'; }
+
+# The slugs belonging to $1: its own, plus one per live git worktree nested
+# underneath it. One per line, sorted, no duplicates.
+#
+# Derived from real worktree paths rather than from the slug string, and that
+# is the whole point of the function. Slugs cannot distinguish a child from a
+# sibling: `/w/proj/wt1` and `/w/proj-wt1` both slug to `-w-proj-wt1`, so a
+# `$slug-*` glob matches any sibling project whose directory name merely
+# extends this one's - on a host with `proj` and `proj-ci` checked out side by
+# side, `proj-ci` reads as part of `proj`. A path test cannot make that
+# mistake, because a sibling is not under the workspace directory.
+#
+# Listing from a worktree names the main checkout and its siblings too. Those
+# are not under $1, so they drop out here: launching against a worktree
+# exposes that worktree and nothing else, which is correct.
+workspace_slugs() {
+    local ws="$1" wt
+    ws="${ws%/}"
+    ws_slug "$ws"
+    printf '\n'
+    # A path containing a newline splits into two lines that name no real
+    # directory, so both fail the -d test below rather than being parsed as
+    # some other worktree. git has no unambiguous non-`-z` output here.
+    git -C "$ws" worktree list --porcelain 2>/dev/null \
+        | sed -n 's/^worktree //p' \
+        | while IFS= read -r wt; do
+            case "$wt" in
+                "$ws"/*) ;;
+                *) continue ;;
+            esac
+            [ -d "$wt" ] || continue
+            ws_slug "$wt"
+            printf '\n'
+        done
+}
+
+# Every slug in workspace_slugs, sorted and deduplicated.
+workspace_slugs_sorted() { workspace_slugs "$1" | sort -u; }
+
+# The Claude Code session ids belonging to workspace $2, one per line, taken
+# from the transcript filenames under $1 (a ~/.claude/projects directory).
+#
+# ~/.claude/file-history is keyed by session id rather than by project, so
+# there is no slug in its own paths to scope it by. The transcript is what
+# ties the two together: projects/<slug>/<id>.jsonl and file-history/<id>
+# name the same session, so the slugs this workspace owns select the ids it
+# owns. A session id that has no transcript belongs to no project this can
+# see, and is left out.
+workspace_session_ids() {
+    local projects_dir="$1" ws="$2" slug f id
+    [ -d "$projects_dir" ] || return 0
+    while IFS= read -r slug; do
+        [ -n "$slug" ] || continue
+        [ -d "$projects_dir/$slug" ] || continue
+        for f in "$projects_dir/$slug"/*.jsonl; do
+            [ -f "$f" ] || continue
+            id="$(basename "$f" .jsonl)"
+            # Only a real session id, never whatever else a directory might
+            # hold: this value becomes a mount path.
+            case "$id" in
+                *[!a-fA-F0-9-]*) continue ;;
+            esac
+            [ -n "$id" ] && printf '%s\n' "$id"
+        done
+    done < <(workspace_slugs_sorted "$ws") | sort -u
+}

bin/dejima-ttyd

@@ -5,37 +5,27 @@
 # systemd/dejima-ttyd.service.
 # See docs/remote-access.md.
 #
-# --interface 127.0.0.1 is a security requirement, not a preference.
-# docker/init-firewall.sh:187 accepts container OUTPUT to the default-route
-# address - the Docker bridge gateway, i.e. the host - and attaching tmux
-# gives a HOST shell (the prefix key opens a new window outside the
-# container). A listener on any address the container can route to would hand
-# a compromised session its own escape hatch. Loopback is not one of those
-# addresses: the container has no route to the host's 127.0.0.1. Put nginx in
-# front, bound to one specific LAN address - see
-# docs/nginx-dejima.conf.example.
+# --interface 127.0.0.1 is a requirement, not a preference. init-firewall.sh
+# accepts container OUTPUT to the bridge gateway (the host), and attaching
+# tmux gives a HOST shell - the prefix key opens a window outside the
+# container. A listener on any address the container can route to is a
+# compromised session's escape hatch; loopback is not one of them. Put nginx
+# in front, bound to one LAN address - docs/nginx-dejima.conf.example.
 #
-# ttyd is given no credential of its own, so the reverse proxy is the only
-# thing authenticating a request, and the loopback bind is what makes that
-# safe rather than careless: nothing but a process on this host can open the
-# socket at all. --writable is what makes the terminal a terminal rather than
-# a viewer.
+# ttyd carries no credential of its own, so the proxy is the only thing
+# authenticating and the loopback bind is what makes that safe. --writable
+# makes it a terminal rather than a viewer.
 #
-# --check-origin is the other half of that, and it is not optional. Without
-# it ttyd does no origin validation at all (it only calls check_host_origin
-# when the flag is set), and browsers do not apply the same-origin policy to
-# WebSockets: any page the operator's browser loads could open
-# wss://<front-door>/ws, have the browser attach its cached credential for
-# that origin, and be handed this host shell. Proxy authentication cannot
-# stop that - it is satisfied by the victim's own browser. ttyd compares the
-# Origin authority against the Host header, which is why the nginx example
-# passes Host through unchanged; the two go together.
+# --check-origin is not optional: without it ttyd does no origin validation
+# at all, and browsers do not apply same-origin to WebSockets, so any page
+# the operator loads could open wss://<front-door>/ws with the browser
+# attaching its cached credential and be handed this shell. Proxy auth cannot
+# stop that - the victim's own browser satisfies it. ttyd compares Origin
+# against Host, which is why the nginx example passes Host through unchanged.
 #
-# --url-arg lets the client pass ?arg=<run-id>, which is what the notification
-# deep link uses. It reaches dejima-attach as argv, never as a shell word, and
-# anything there that is not --list or the empty string (a bare `?arg=`, which
-# lands on the picker) has to match the run-id shape before it is used as a
-# path component or anything else.
+# --url-arg carries ?arg=<run-id> for the notification deep link. It reaches
+# dejima-attach as argv, never a shell word, and anything but --list or empty
+# must match the run-id shape before it is used as a path component.
 set -euo pipefail
 
 SCRIPT_DIR="$(cd "$(dirname "$(readlink -f "${BASH_SOURCE[0]}")")" && pwd)"

bin/harness/claude-code.sh

@@ -0,0 +1,232 @@
+# Harness plugin: Claude Code.
+#
+# Sourced by bin/dejima through harness_plugin_load when [core] harness names
+# this harness. Part of the trusted computing base - see the interface notes
+# in bin/dejima-harness-lib.sh for why plugins live here and why they are
+# sourced rather than invoked.
+#
+# What is here is what could not be stated as a manifest value. The paths
+# themselves come from the manifest ([harness.claude-code] projects_dir); the
+# reason this is code is the pairing in the second block, which resolves one
+# directory tree through the filenames of another.
+
+# Registers the waiting/done notification handlers into dejima's settings
+# layer, in the shape Claude Code reads them.
+#
+# Claude Code merges hook entries across settings layers rather than letting
+# the higher one replace the lower, so these are added to whatever the
+# operator already has rather than displacing it. $HOME is left unexpanded
+# for Claude Code to resolve at hook-run time - inside the container, not
+# here - matching the convention install.sh already uses.
+#
+# Written through a temp file and renamed: the destination is the settings
+# layer bin/dejima is about to mount, and a jq that failed partway would
+# otherwise leave the session with a truncated one.
+harness_register_notify() {
+    local settings="$1" tmp
+    tmp="$(mktemp "${TMPDIR:-/tmp}/dejima-settings-remote.XXXXXX.json")"
+    if jq '.hooks.Notification = ((.hooks.Notification // []) + [{"matcher": "", "hooks": [{"type": "command", "command": "$HOME/.claude/hooks/dejima-notify.sh waiting"}]}])
+      | .hooks.Stop = ((.hooks.Stop // []) + [{"matcher": "", "hooks": [{"type": "command", "command": "$HOME/.claude/hooks/dejima-notify.sh done"}]}])' \
+        "$settings" > "$tmp"; then
+        mv "$tmp" "$settings"
+    else
+        rm -f "$tmp"
+        echo "warning: could not register the notification hooks; remote access" >&2
+        echo "will not be told when this session wants attention." >&2
+    fi
+}
+
+# The per-project and per-session state this harness keeps on the host, and
+# which parts of it this workspace is entitled to see.
+harness_state_mounts() {
+    local ws="$1" projects slug sid
+    projects="$(harness_config_get projects_dir)"
+    [ -n "$projects" ] || return 0
+
+    # projects/, one slug at a time: the tree is every project's transcripts
+    # and memory, and memory is loaded into context at session start, so a
+    # write into another slug is background a future session there believes
+    # (#104). projects/ itself stays container-local, so an unmounted slug is
+    # still writable and just does not survive - which is how Claude Code
+    # creates a slug it does not find. Existing slugs only, apart from the
+    # workspace's own, so a checkout full of worktrees does not litter the
+    # host with empty directories for worktrees nothing has run in.
+    mkdir -p "$projects/$(ws_slug "$ws")"
+    while IFS= read -r slug; do
+        [ -n "$slug" ] || continue
+        [ -d "$projects/$slug" ] || continue
+        emit_mount "$projects/$slug"
+    done < <(workspace_slugs_sorted "$ws")
+
+    # file-history, same shape, sharper reason: `<session>/<hash>@v1` is a
+    # verbatim copy of an edited file, so the tree is other projects' source
+    # rather than a record of it, and it was writable - tampering there puts
+    # chosen content into a later restore. Keyed by session id, so there is
+    # no slug to match on; workspace_session_ids resolves it through the
+    # transcript names. The starting session's id cannot be known yet, so its
+    # directory is container-local: undo history for it is lost if it is
+    # resumed on the host.
+    #
+    # The pairing is why the plugin is code, and the split is the design in
+    # miniature: both directories are statable, so both are manifest keys.
+    # What cannot be stated is the rule that a name in one selects an entry
+    # in the other, so that is here.
+    local history
+    history="$(harness_config_get file_history_dir)"
+    [ -n "$history" ] || return 0
+    mkdir -p "$history"
+    while IFS= read -r sid; do
+        [ -n "$sid" ] || continue
+        [ -d "$history/$sid" ] || continue
+        emit_mount "$history/$sid"
+    done < <(workspace_session_ids "$projects" "$ws")
+}
+
+# Puts dejima's guard hooks where Claude Code will find them: symlinked into
+# ~/.claude/hooks, and registered in ~/.claude/settings.json. Run by
+# install.sh, not at launch - this writes the operator's own config, and the
+# operator is who decides to install.
+#
+# $1 is the repo root, $2 the directory to move a displaced file into. Every
+# change is printed as it happens, because it is someone's real config.
+#
+# Each guard is gated on the harness having declared it: a harness declares
+# what it can host (see `guards` in its manifest), and registering something
+# it did not declare would contradict the launch message that says so.
+harness_register_guards() {
+    local repo="$1" backup="$2" settings="$HOME/.claude/settings.json"
+
+    mkdir -p "$HOME/.claude/hooks"
+
+    _link_hook() {
+        local name="$1" src="$repo/hooks/$1" dest="$HOME/.claude/hooks/$1"
+        if [ -L "$dest" ] && [ "$(readlink -f "$dest")" = "$(readlink -f "$src")" ]; then
+            echo "Skipping: $name already linked to this repo's copy"
+            return
+        fi
+        if [ -e "$dest" ] || [ -L "$dest" ]; then
+            mkdir -p "$backup"
+            echo "Moving: existing ~/.claude/hooks/$name to $backup/$name"
+            mv "$dest" "$backup/$name"
+        fi
+        echo "Linking: $name ($src -> $dest)"
+        ln -s "$src" "$dest"
+    }
+
+    # Not a guard: this is R3, the briefing's delivery mechanism, and a
+    # session without it does not know it is in a sandbox.
+    _link_hook sandbox-context.sh
+    harness_has_guard push-guard  && _link_hook git-push-guard.sh
+    harness_has_guard merge-guard && _link_hook merge-guard.sh
+
+    # Symlink from before the gitea-merge-guard.sh -> merge-guard.sh rename
+    # points at a hooks/ path that no longer exists. Only remove it if it is
+    # actually dangling, so an operator's own file at this name is left alone.
+    local old_merge_guard="$HOME/.claude/hooks/gitea-merge-guard.sh"
+    if [ -L "$old_merge_guard" ] && [ ! -e "$old_merge_guard" ]; then
+        echo "Removing: dangling pre-rename symlink $old_merge_guard"
+        rm -f "$old_merge_guard"
+    fi
+
+    # Linked so it exists at a stable path for the read-only ~/.claude/hooks
+    # mount, but deliberately NOT registered below: bin/dejima registers it
+    # per session, into a throwaway settings copy, and only when remote
+    # access is configured. Registering it here would fire it for every
+    # unsandboxed host session too, where it has no marker directory to
+    # write to.
+    harness_has_guard notify && _link_hook dejima-notify.sh
+
+    # deny-dangerous-bash.sh is the one hook that is NOT sandbox-gated: it
+    # has no DEJIMA check, so it applies to the operator's own host sessions
+    # as well. That is deliberate, and it is why it may already be owned by
+    # another install (dotfiles, say) - only claim it if nothing else
+    # provides it, so this repo works standalone without clobbering an
+    # existing setup.
+    if harness_has_guard bash-deny; then
+        local deny_dest="$HOME/.claude/hooks/deny-dangerous-bash.sh"
+        if [ -e "$deny_dest" ] || [ -L "$deny_dest" ]; then
+            echo "Skipping: deny-dangerous-bash.sh already present at $deny_dest (leaving it as-is)"
+        else
+            echo "Linking: deny-dangerous-bash.sh ($repo/hooks/deny-dangerous-bash.sh -> $deny_dest)"
+            ln -s "$repo/hooks/deny-dangerous-bash.sh" "$deny_dest"
+        fi
+    fi
+    echo
+
+    if [ ! -f "$settings" ]; then
+        echo "Creating $settings"
+        echo '{}' > "$settings"
+    fi
+
+    # The registration that went with the symlink removed above. Two things
+    # depend on it going: Claude Code reports a missing hook command on every
+    # session start while it points at a path that no longer exists, and -
+    # until the dedupe below stopped matching it - it also stood in for the
+    # registration it was renamed from, so `_register merge-guard.sh` skipped
+    # and the merge guard was registered nowhere at all. Dropped only when
+    # the path it names is gone, so an operator's own file at that name keeps
+    # its registration, matching the symlink check above.
+    if [ ! -e "$old_merge_guard" ] && jq -e --arg s "/gitea-merge-guard.sh" \
+        '[.hooks.PreToolUse[]?.hooks[]?.command // empty] | any(endswith($s))' \
+        "$settings" >/dev/null 2>&1; then
+        echo "Removing: stale pre-rename gitea-merge-guard.sh registration from $settings"
+        local stale_tmp; stale_tmp=$(mktemp)
+        jq --arg s "/gitea-merge-guard.sh" \
+            '.hooks.PreToolUse = ((.hooks.PreToolUse // [])
+                | map(.hooks = ((.hooks // []) | map(select((.command // "") | endswith($s) | not))))
+                | map(select((.hooks | length) > 0)))' \
+            "$settings" > "$stale_tmp" && mv "$stale_tmp" "$settings"
+    fi
+
+    _register() {
+        local event="$1" matcher="$2" script_name="$3" description="$4"
+        # endswith on the leading slash, not test(): test() is an unanchored
+        # regex, so a hook whose name merely ends in another's - which is what
+        # a rename like gitea-merge-guard.sh -> merge-guard.sh produces - read
+        # as already registered, and the new hook was silently never added.
+        if jq -e --arg s "/$script_name" \
+            "[.hooks.${event}[]?.hooks[]?.command // empty] | any(endswith(\$s))" \
+            "$settings" >/dev/null 2>&1; then
+            echo "Skipping: $script_name already registered as a $event hook"
+            return
+        fi
+        echo "Registering $script_name as a $event hook in $settings ($description)"
+        local tmp
+        tmp=$(mktemp)
+        jq --arg matcher "$matcher" --arg cmd "\$HOME/.claude/hooks/$script_name" \
+            ".hooks.${event} = ((.hooks.${event} // []) + [{\"matcher\": \$matcher, \"hooks\": [{\"type\": \"command\", \"command\": \$cmd}]}])" \
+            "$settings" > "$tmp" && mv "$tmp" "$settings"
+    }
+
+    _register SessionStart "startup|resume|clear|compact" sandbox-context.sh \
+        "injects sandbox context on session start"
+    harness_has_guard push-guard && _register PreToolUse Bash git-push-guard.sh \
+        "blocks pushes outside the allowed host(s)/branches"
+    harness_has_guard merge-guard && _register PreToolUse Bash merge-guard.sh \
+        "blocks the configured forge's merge command inside the sandbox"
+    harness_has_guard bash-deny && _register PreToolUse Bash deny-dangerous-bash.sh \
+        "blocks obviously destructive Bash commands"
+    echo
+    return 0
+}
+
+# What Claude Code causes to run on the HOST, unsandboxed.
+#
+# settings.json's statusLine.command names a script the host executes on
+# every session, so a session that can write that file has a path onto the
+# host that no container boundary sees. bin/dejima stages the file rather
+# than mounting over it, because a :ro mount landing on a symlink - the
+# common case - protects the target and leaves the link replaceable;
+# detecting a change here is the compensating control for that.
+#
+# The namespace and leaf together form the id a baseline stores
+# ("statusline:statusline.sh"), so neither may change while anyone has
+# approvals recorded. The label is the fixed prefix `ack` strips to recover
+# that id, which is why it lives here as a literal rather than in the
+# manifest - see the host-exec notes in bin/dejima-harness-lib.sh.
+harness_host_exec_surfaces() {
+    local p
+    p="$(harness_config_get statusline_path)"
+    emit_host_exec statusline statusline.sh "$p" \
+        "HOST-CODE-EXECUTION statusline.sh changed (runs UNSANDBOXED on the host via settings.json statusLine.command): "
+}

docker/Dockerfile

@@ -77,19 +77,15 @@ ENV DEJIMA=1
 # this user, so it doesn't need to own anything itself.
 RUN useradd -r -s /usr/sbin/nologin ulogd
 
-# Every fetch from here down retries, and none is single-shot. What matters
-# isn't the odds of a particular download failing, it's the odds of any of
-# them failing - and without retries a single blip anywhere throws away the
-# whole build.
+# Every fetch from here down retries. What matters is not the odds of one
+# download failing but of ANY of them failing, and without retries a single
+# blip discards the whole build.
 #
-# These steps deliberately sit *above* the SANDBOX_* ARGs, not below. Nothing
-# between here and `USER` needs the sandbox user to exist yet - they all
-# install system-wide as root - so keeping them ahead of the ARGs means a
-# build that overrides one (test/runtime.bats does, and so does any host
-# whose uid isn't 1000) still reuses every layer down to this point instead
-# of re-running all of it. That single ordering was most of the cost of the
-# override rebuild: it was 1169s of a 1379s test run, for a difference that
-# amounts to one useradd.
+# These steps sit ABOVE the SANDBOX_* ARGs on purpose: nothing here needs the
+# sandbox user, so a build overriding a uid (runtime.bats, and any host not
+# on 1000) still reuses every layer to this point. That ordering alone was
+# most of the cost of an override rebuild, for a difference amounting to one
+# useradd.
 # gosu lets entrypoint.sh (root) drop to the node user for the actual
 # session after setting up the firewall, without leaving any setuid/sudo
 # escalation path available to node afterward. Verified against the
@@ -185,7 +181,11 @@ RUN mkdir /commandhistory && \
 # from the host, and Docker creates the parents of a mount point root-owned
 # if they don't already exist - which would leave the sandbox user unable to
 # write the caches and history it keeps directly in ~/.claude.
-RUN mkdir -p $SANDBOX_HOME/.ssh $SANDBOX_HOME/.claude && \
+# projects/ is in that list for the same reason one level down: it is no
+# longer mounted itself, only individual slug directories inside it, so
+# without this Docker creates it root-owned while setting those up and the
+# session cannot write a slug of its own.
+RUN mkdir -p $SANDBOX_HOME/.ssh $SANDBOX_HOME/.claude $SANDBOX_HOME/.claude/projects $SANDBOX_HOME/.claude/file-history && \
   chown -R $SANDBOX_USER:$SANDBOX_USER $SANDBOX_HOME/.ssh $SANDBOX_HOME/.claude && \
   chmod 700 $SANDBOX_HOME/.ssh $SANDBOX_HOME/.claude
 
@@ -255,24 +255,23 @@ RUN npm install -g tree-sitter-cli@0.24.7
 # matching Chromium build into node's own cache, both at build time so a
 # session never needs network access for either.
 #
-# Chromium's own internal process sandbox is questionable here: it launches
-# without --no-sandbox and without error, but its setuid sandbox helper
-# (chrome_sandbox) isn't actually setuid-root in this image, and this host
-# has the same unprivileged-userns restriction that broke bubblewrap - so
-# its namespace-based isolation almost certainly isn't active either, it's
-# just not hard-failing the way bubblewrap did. Not treated as a real
-# security boundary either way: this container is already the boundary that
-# matters, and a compromised session already has full `node`-level bash
-# access, which a Chromium exploit couldn't exceed regardless of whichever
-# of its own internal sandbox layers happen to be active.
+# Chromium's own process sandbox is probably not active here: it launches
+# without error, but chrome_sandbox is not setuid-root in this image and the
+# unprivileged-userns restriction that broke bubblewrap applies, so it just
+# fails softer than bubblewrap did. Not treated as a boundary either way -
+# the container is the boundary, and a session already has node-level shell
+# access that a Chromium exploit could not exceed.
 RUN npm install -g playwright@${PLAYWRIGHT_VERSION} && \
   npx playwright install chromium
 
 # Copy and set up firewall + DNS-proxy + entrypoint scripts, every forge's
-# CLI helpers, and the claude wrapper that pins
-# --dangerously-skip-permissions. The launched application itself
-# (SANDBOX_CMD, default `claude`) and any editor config for it are supplied
-# at runtime, not baked into the image - see docker-compose.yml/entrypoint.sh.
+# CLI helpers, and the harness wrapper that pins whichever no-confirm flags
+# the selected harness needs. The wrapper is installed under a neutral name:
+# the image is shared across harnesses and the selection is per launch, so
+# entrypoint.sh is what links it into place as the harness's own command.
+# The launched application itself (SANDBOX_CMD) and any editor config for it
+# are supplied at runtime, not baked into the image - see
+# docker-compose.yml/entrypoint.sh.
 COPY docker/fixed-domains.sh /usr/local/bin/
 COPY docker/dns-proxy.sh /usr/local/bin/
 COPY docker/init-firewall.sh /usr/local/bin/
@@ -282,7 +281,7 @@ COPY docker/ulogd.conf /etc/ulogd-sandbox.conf
 COPY docker/sandbox-clipboard.lua /usr/local/share/nvim/runtime/plugin/sandbox-clipboard.lua
 COPY docker/entrypoint.sh /usr/local/bin/
 COPY docker/anthropic-proxy-forwarder.js /usr/local/bin/anthropic-proxy-forwarder.js
-COPY docker/claude-wrapper.sh /usr/local/bin/claude
+COPY docker/harness-wrapper.sh /usr/local/bin/dejima-harness-wrapper
 # Back to root before the first RUN that writes outside /home/node. The npm
 # installs above run as node deliberately; COPY writes as root regardless of
 # USER, so the copies above are unaffected, but the install(1) below is a RUN
@@ -296,7 +295,7 @@ COPY forges/ /tmp/forges/
 RUN find /tmp/forges -type f -path '*/bin/*' -exec install -m 755 {} /usr/local/bin/ \; \
     && rm -rf /tmp/forges
 
-RUN chmod +x /usr/local/bin/dns-proxy.sh /usr/local/bin/init-firewall.sh /usr/local/bin/log-shipper.sh /usr/local/bin/entrypoint.sh /usr/local/bin/claude && \
+RUN chmod +x /usr/local/bin/dns-proxy.sh /usr/local/bin/init-firewall.sh /usr/local/bin/log-shipper.sh /usr/local/bin/entrypoint.sh /usr/local/bin/dejima-harness-wrapper && \
   echo 'export PATH="$PATH:/usr/local/share/npm-global/bin"' > /etc/profile.d/npm-global-path.sh
 
 # Entrypoint runs as root (no USER node here) so it can set up the firewall

docker/claude-wrapper.sh

@@ -1,16 +0,0 @@
-#!/bin/bash
-# Shadows the real `claude` binary (installed under npm-global/bin, later on
-# PATH) so every invocation - interactive shell, or claudecode.nvim's
-# terminal - runs with --dangerously-skip-permissions, without callers
-# needing to know that.
-#
-# The sandbox's settings overrides arrive the same way: as an extra
-# --settings layer rather than by mounting over ~/.claude/settings.json (see
-# docker-compose.yml and #47). That only holds if every entry point passes
-# the flag, and shadowing the binary is what makes it true for the
-# invocations the entrypoint never sees. The file is absent when a container
-# is started without bin/dejima, as test/runtime.bats does.
-settings=()
-[ -f /etc/dejima-settings.json ] && settings=(--settings /etc/dejima-settings.json)
-
-exec /usr/local/share/npm-global/bin/claude --dangerously-skip-permissions "${settings[@]}" "$@"

docker/docker-compose.yml

@@ -68,9 +68,18 @@ services:
       # just live at ~/.gitconfig directly.
       GIT_CONFIG_GLOBAL: ${HOME}/.gitconfig-sandbox/.gitconfig
       ALLOWED_DOMAINS: ${ALLOWED_DOMAINS:-}
-      # What actually runs as the sandbox user - defaults to bare `claude`.
-      # See entrypoint.sh and bin/dejima's SANDBOX_NVIM_CONFIG.
+      # What actually runs as the sandbox user - the selected harness's own
+      # command unless [core] command overrides it. See entrypoint.sh and
+      # bin/dejima's SANDBOX_NVIM_CONFIG.
       SANDBOX_CMD: ${SANDBOX_CMD:-claude}
+      # The selected harness's manifest, resolved host-side: which command
+      # the wrapper shadows, the real binary behind it, the flags that make
+      # the session unattended, and how this harness is told to read dejima's
+      # settings layer. Nothing about a harness is baked into the image.
+      SANDBOX_HARNESS_COMMAND: ${SANDBOX_HARNESS_COMMAND:-}
+      SANDBOX_HARNESS_REAL: ${SANDBOX_HARNESS_REAL:-}
+      SANDBOX_HARNESS_ARGS: ${SANDBOX_HARNESS_ARGS:-}
+      SANDBOX_HARNESS_SETTINGS: ${SANDBOX_HARNESS_SETTINGS:-}
       # Opt-in escape hatch from the default hard GitHub-push block - see
       # init-firewall.sh and git-push-guard.sh. Off unless explicitly set.
       ALLOW_GITHUB_PUSH: ${ALLOW_GITHUB_PUSH:-0}
@@ -111,33 +120,24 @@ services:
       # (based on cwd) lines up with the host, letting a session started
       # here be resumed on the host with `claude --resume`.
       - ${WORKSPACE_DIR:?set WORKSPACE_DIR to the project directory to mount}:${WORKSPACE_DIR}
-      # ~/.claude is deliberately not mounted as a whole. It used to be,
-      # read-write, with the config inside it layered back read-only on top.
-      # Every one of those overlays was a mount point nested in a
-      # host-writable mount, and a single-file bind mount is pinned to the
-      # dentry at that path: a host-side rename of the underlying file - a
-      # sync daemon, or Claude Code's own atomic writes - unlinks it, the
-      # mount silently detaches, and the path falls through to the real,
-      # writable host file (#47, #48). The same thing happened to the mask
-      # over .credentials.json, handing the sandbox the operator's real
-      # OAuth token on the first refresh (#50).
+      # ~/.claude is enumerated, never mounted whole. A single-file bind
+      # mount is pinned to the dentry at its path, so a host-side rename -
+      # a sync daemon, or Claude Code's own atomic writes - unlinks it, the
+      # mount detaches silently, and the path falls through to the real
+      # writable host file (#47, #48; #50 did exactly that to the
+      # .credentials.json mask). Anything not listed here is
+      # container-local, which is the safer default but means new state that
+      # should persist has to be added deliberately.
       #
-      # So the contents are enumerated instead. Anything not listed below is
-      # container-local: writable during the session, gone with the
-      # container. That is the safer default - a config file a future Claude
-      # Code version adds simply doesn't persist, instead of sitting
-      # unprotected on a writable mount - but it does mean new state that
-      # *should* persist has to be added here deliberately.
+      # Read-write below, and all directories: a directory mount can't
+      # detach that way, because a rename replaces an entry within it rather
+      # than the mount point itself.
       #
-      # Runtime state that has to survive the session, read-write. All
-      # directories, which is what makes them safe: a directory bind mount
-      # can't detach the way a file one does, because a rename replaces an
-      # entry *within* the mounted directory rather than the mount point.
-      # projects/ is what host-side `claude --resume` reads (and what
-      # log-shipper.sh tails); the rest is Claude Code's edit-undo record
-      # and the task/plan state skills write.
-      - ${HOME}/.claude/projects:${HOME}/.claude/projects
-      - ${HOME}/.claude/file-history:${HOME}/.claude/file-history
+      # projects/ and file-history/ are absent on purpose - the trees are
+      # every project's memory and every project's file contents, and a
+      # write into either reaches sessions elsewhere (#104). bin/dejima
+      # mounts this workspace's own entries, one -v each; see
+      # bin/dejima-slug-lib.sh for how that set is chosen.
       - ${HOME}/.claude/backups:${HOME}/.claude/backups
       - ${HOME}/.claude/tasks:${HOME}/.claude/tasks
       - ${HOME}/.claude/plans:${HOME}/.claude/plans
@@ -170,7 +170,7 @@ services:
       - ${HOME}/.claude/CLAUDE.md:/etc/dejima-claude/CLAUDE.md:ro
       # The sandbox's own settings overrides (sandbox.enabled off, plus the
       # remote-access notify hooks when that's on), generated per run by
-      # bin/dejima and reaching Claude Code as claude-wrapper.sh's
+      # bin/dejima and reaching the harness as harness-wrapper.sh's
       # --settings argument. A container-local path, like the staged config
       # above and for the same reason (#47) - and separate from that staging
       # directory because this file is generated per run rather than copied
@@ -233,6 +233,18 @@ services:
 
 volumes:
   claude-code-bashhistory:
+  # Named explicitly so bin/dejima can give each project its own set. These
+  # persist across containers by design (a plugin install should not have to
+  # happen every launch), which without the split meant one project's session
+  # could leave a Lua file or a mason/bin binary that another project's
+  # session executes the next time nvim runs. The service still refers to the
+  # static keys; only the volume Docker actually creates varies.
+  #
+  # The defaults reproduce the previous names exactly, so a `docker compose`
+  # invocation that bypasses bin/dejima behaves as it did.
   nvim-data:
+    name: ${SANDBOX_NVIM_DATA:-dejima_nvim-data}
   nvim-state:
+    name: ${SANDBOX_NVIM_STATE:-dejima_nvim-state}
   nvim-cache:
+    name: ${SANDBOX_NVIM_CACHE:-dejima_nvim-cache}

docker/entrypoint.sh

@@ -103,6 +103,30 @@ if [ -s "$SANDBOX_SSH_KEY_SRC" ]; then
     export SSH_AUTH_SOCK
 fi
 
+# Put the harness wrapper on PATH under the harness's own command name,
+# ahead of the real binary, so every invocation gets the no-confirm flags -
+# including the ones dejima never sees, an interactive shell or an editor
+# plugin's terminal. Linked here rather than COPYed at build time because
+# the image is shared across harnesses and the selection is per launch.
+#
+# Root-owned in a root-owned directory: the sandbox user must not be able to
+# replace the wrapper with something that drops the flags, or reinstate the
+# real binary's name ahead of it.
+#
+# Skipped when SANDBOX_HARNESS_COMMAND is unset, which is a container started
+# without bin/dejima (test/runtime.bats). Refused when it is set to anything
+# but a bare command name: it becomes a path under /usr/local/bin, and a
+# value with a slash in it would write somewhere else entirely.
+if [ -n "${SANDBOX_HARNESS_COMMAND:-}" ]; then
+    case "$SANDBOX_HARNESS_COMMAND" in
+        */*|.*|'')
+            echo "entrypoint: refusing harness command name: $SANDBOX_HARNESS_COMMAND" >&2
+            exit 1 ;;
+    esac
+    ln -sf /usr/local/bin/dejima-harness-wrapper \
+        "/usr/local/bin/$SANDBOX_HARNESS_COMMAND"
+fi
+
 # ~/.claude config files arrive staged read-only under /etc/dejima-claude
 # (see the mount inventory in docker-compose.yml) and get copied into place
 # here rather than bind-mounted at their real paths. A copy can't be
@@ -252,27 +276,18 @@ fi
 [ -n "${ANTHROPIC_BASE_URL:-}" ] || unset ANTHROPIC_BASE_URL
 [ -n "${ANTHROPIC_AUTH_TOKEN:-}" ] || unset ANTHROPIC_AUTH_TOKEN
 
-# Remote access is opt-in and off by default, so this whole block is skipped
-# for the sessions that never asked for it: DEJIMA_NOTIFY_DIR is set only by
-# bin/dejima's remote_enabled branch. An unset value is the ordinary case and
-# says nothing.
+# Unset is the ordinary case (remote access is opt-in) and says nothing.
+# Set-but-unusable is worth a startup line because of how it presents: the
+# marker directory's source is created per launch and removed on exit, so
+# restarting the container rather than relaunching `dejima` leaves it gone,
+# and Docker answers a missing bind source by creating one root-owned. The
+# hook's `[ -d ]` guard then passes and every notification fails EACCES,
+# while the session is also missing from `dejima-attach` because its registry
+# entry went with the same directory - three unrelated-looking symptoms.
 #
-# Set-but-unusable is different, and it is worth a line at startup because of
-# how it otherwise presents. The marker directory is a bind mount whose source
-# bin/dejima creates per launch and removes on exit, so restarting the
-# container - `docker start`, or a restart policy - rather than relaunching
-# `dejima` leaves the source gone. Docker's answer to a missing bind source is
-# not to refuse: it creates one, root-owned. The directory then exists, so the
-# notify hook's own `[ -d ]` guard passes and every notification fails with
-# EACCES instead; the session is also absent from `dejima-attach`, because its
-# registry entry went with the same directory. Three unrelated-looking symptoms
-# and no obvious moment they started.
-#
-# Reported, not repaired. Chowning this would restore notifications while the
-# session stayed missing from the registry - which bin/dejima writes host-side
-# and never mounts here - so a session would look fixed while half of the
-# feature was still broken. Relaunching is the actual remedy, and it is one the
-# message can name.
+# Reported, not repaired: chowning would restore notifications while the
+# session stayed absent from the registry, looking fixed with half the
+# feature broken. Relaunching is the remedy, and the message can say so.
 if [ -n "${DEJIMA_NOTIFY_DIR:-}" ]; then
     if [ ! -d "$DEJIMA_NOTIFY_DIR" ]; then
         echo "WARNING: remote access is enabled, but $DEJIMA_NOTIFY_DIR is not a directory." >&2
@@ -301,10 +316,8 @@ fi
 # sudo/setuid path back to root after this, so it can't re-run
 # init-firewall.sh itself to widen the allowlist mid-session.
 #
-# SANDBOX_CMD is what actually runs as that user - defaults to bare `claude`
-# (the wrapped binary on PATH, see claude-wrapper.sh) so a fresh clone of
-# this repo works with no editor setup at all. Override it (e.g. to `nvim`)
-# for a custom editor-driven workflow; pair with SANDBOX_NVIM_CONFIG (see
-# bin/dejima) to bring in a real config for it instead of running
-# unconfigured.
+# SANDBOX_CMD is what actually runs as that user - the selected harness's
+# own command unless [core] command overrides it (e.g. to `nvim`) for an
+# editor-driven workflow; pair that with SANDBOX_NVIM_CONFIG (see bin/dejima)
+# to bring in a real config for it instead of running unconfigured.
 exec gosu "$SANDBOX_USER" ${SANDBOX_CMD:-claude} "$@"

docker/fixed-domains.sh

@@ -50,22 +50,17 @@ GITHUB_DNS_SUFFIXES=(
 # Which ipset an allowlisted name's addresses belong in.
 #
 # GitHub's are restricted to port 443 by init-firewall.sh, and that
-# restriction is the read-only guarantee - there is no transport left for the
-# forwarded ssh-agent's key to push with. allowed-domains carries no port
-# restriction at all, so an address that reaches it is reachable on 22 as
-# well, and a GitHub address that reaches it silently undoes the whole
-# arrangement.
+# restriction IS the read-only guarantee: no transport is left for a
+# forwarded agent key to push with. allowed-domains has no port restriction,
+# so a GitHub address landing there silently undoes the arrangement.
 #
-# dns-proxy.sh gets this right for free: dnsmasq applies the longest matching
-# ipset= line, so GitHub names go to github-domains however they were
-# allowlisted. Every other path has to ask, and two did not - the
-# extra-domains loop in init-firewall.sh and the live add in bin/dejima-allow
-# both added unconditionally to allowed-domains. A project with github.com in
-# its .dejima/allowed-domains therefore opened port 22 to whichever address
-# GitHub returned at container start, for as long as that answer stayed
-# cached. It presented as an intermittent startup failure, because the
-# port-22 self-check re-resolves the name and only sees the hole when the
-# second answer matches the first.
+# dnsmasq applies the longest matching ipset= line, so dns-proxy.sh gets this
+# right for free. Every other path has to ask, and two did not - the
+# extra-domains loop and bin/dejima-allow's live add both appended
+# unconditionally. A project declaring github.com therefore opened port 22 to
+# whatever address was cached at container start, presenting as an
+# intermittent startup failure because the self-check re-resolves and only
+# sees the hole when the second answer matches the first.
 ipset_for_domain() {
     local domain="$1" suffix
     for suffix in "${GITHUB_DNS_SUFFIXES[@]}"; do

docker/harness-wrapper.sh

@@ -0,0 +1,51 @@
+#!/bin/bash
+# Shadows the selected harness's real binary on PATH, so every invocation -
+# the entrypoint's, an interactive shell's, an editor plugin's terminal -
+# runs with the flags dejima requires, without any caller needing to know
+# that.
+#
+# Installed by entrypoint.sh under the harness's own command name rather than
+# COPYed to a fixed one at build time: the image is shared across harnesses
+# and the selection is per launch, so which name to shadow is only known once
+# the container starts.
+#
+# Every value here comes from the selected harness's manifest, resolved on
+# the host (see bin/dejima) and passed in as environment. Nothing about any
+# one harness is written down in this file.
+#
+#   SANDBOX_HARNESS_REAL      the real binary this shadows
+#   SANDBOX_HARNESS_ARGS      the no-confirm flags, without which the agent
+#                             stops to ask and the arrangement becomes a
+#                             worse way of running it on the host
+#   SANDBOX_HARNESS_SETTINGS  how this harness is told to read dejima's own
+#                             settings layer, if it reads one at all
+#
+# The settings layer arrives as an extra layer rather than as a mount over
+# the harness's own config file (docker-compose.yml, #47). That only holds if
+# every entry point passes it, which is what shadowing the binary buys.
+# /etc/dejima-settings.json is absent when a container is started without
+# bin/dejima, as test/runtime.bats does, so its flag stays conditional on the
+# file rather than on the manifest key alone.
+#
+# Word-split deliberately: both carry flags, which have no spaces, and an
+# array is what exec needs. A harness whose flags need quoting has outgrown
+# this and should say so in its plugin.
+set -u
+
+if [ -z "${SANDBOX_HARNESS_REAL:-}" ] || [ ! -x "${SANDBOX_HARNESS_REAL:-}" ]; then
+    echo "dejima: no harness binary at '${SANDBOX_HARNESS_REAL:-<unset>}'." >&2
+    echo "dejima: command_path in the harness manifest is what names it." >&2
+    exit 127
+fi
+
+args=()
+# shellcheck disable=SC2206
+[ -n "${SANDBOX_HARNESS_ARGS:-}" ] && args=(${SANDBOX_HARNESS_ARGS})
+
+settings=()
+# shellcheck disable=SC2206
+[ -n "${SANDBOX_HARNESS_SETTINGS:-}" ] && [ -f /etc/dejima-settings.json ] \
+    && settings=(${SANDBOX_HARNESS_SETTINGS})
+
+exec "$SANDBOX_HARNESS_REAL" \
+    ${args[@]+"${args[@]}"} ${settings[@]+"${settings[@]}"} "$@"

docker/init-firewall.sh

@@ -11,23 +11,17 @@
 set -Eeuo pipefail
 IFS=$'\n\t'
 
-# Nothing in here may die quietly. Under `set -e` a failing command aborts the
-# script with no output at all, and this script's failures are invisible in the
-# worst possible way: the container simply never appears, which reads as the
-# sandbox being broken rather than as one command in one loop having failed.
-# Three CI failures were spent on two different instances of exactly that (#59)
-# - one at the GitHub ranges fetch, one in the domain loop below - and neither
-# left a single line to go on.
+# Nothing here may die quietly. Under `set -e` a failure aborts with no output
+# and the container simply never appears, which reads as the sandbox being
+# broken rather than one command failing. Three CI failures went to two
+# instances of exactly that (#59), neither leaving a line to go on.
 #
-# The explicit `exit 1` paths below say more than this can and are worth
-# keeping; this is the floor under everything else. -E on the line above is
-# what extends it into functions and subshells, where it would otherwise not
-# fire at all.
+# This is the floor; the explicit `exit 1` paths below say more. -E extends it
+# into functions and subshells, where it would not fire at all.
 #
-# A failure inside a command substitution reports twice - once from the
-# subshell, naming the command that actually failed, and once from the
-# statement containing it. That is worth the noise: for `x=$(a | b)` the first
-# line is the only thing that says whether it was `a` or `b`.
+# A failure inside a command substitution reports twice, from the subshell and
+# from the containing statement. Worth the noise: for `x=$(a | b)` the first
+# line is the only thing saying whether it was `a` or `b`.
 trap 'status=$?; echo "ERROR: init-firewall.sh aborted at line $LINENO (exit $status): $BASH_COMMAND" >&2' ERR
 
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
@@ -312,31 +306,23 @@ iptables -P OUTPUT DROP
 iptables -A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT
 iptables -A OUTPUT -m state --state ESTABLISHED,RELATED -j ACCEPT
 
-# Docker's embedded resolver does not answer external names itself - it
-# forwards them to the host's nameservers, and on current Docker it does that
-# from inside THIS network namespace. Those queries therefore traverse the
-# chain below like any other packet, and nothing here can match them: the
-# host's resolver is neither in an ipset nor the bridge gateway. They reach
-# the catch-all REJECT, every lookup comes back SERVFAIL, and the container
-# dies in a self-check that says only that a host was unreachable.
+# Docker's embedded resolver forwards external names to the host's
+# nameservers, and on some versions does it from inside THIS namespace. Those
+# queries then traverse the chain below and match nothing - the host's
+# resolver is neither in an ipset nor the bridge gateway - so they hit the
+# catch-all REJECT and every lookup SERVFAILs. Only the first hop stays on
+# loopback. Two hosts with byte-identical /etc/resolv.conf behaved
+# differently, so this is version-dependent and cannot be assumed either way.
 #
-# The comment on the loopback rule above used to claim DNS "stays on
-# loopback, so this also covers DNS". Only the first hop does. Two hosts with
-# byte-identical /etc/resolv.conf behaved differently here, so this is not a
-# property of any host's configuration and cannot be assumed either way.
+# --uid-owner 0 keeps this from being the hole the DNS allowlist closes: the
+# forwarding socket is dockerd's (root), the session is `node` with no sudo,
+# so it still cannot address a resolver directly. This newly permits DNS only
+# from processes already root in a NET_ADMIN container, which could flush the
+# ruleset anyway.
 #
-# --uid-owner 0 is what keeps this from being the hole the DNS allowlist
-# exists to close. The forwarding socket belongs to dockerd, which is root;
-# the session runs as `node` and there is no sudo in the image, so a
-# compromised session still cannot address a resolver directly and still has
-# no path to a name dns-proxy.sh has not allowed. What this does newly permit
-# is DNS from processes that are already root in a container holding
-# NET_ADMIN - which could simply flush this ruleset instead, so it grants
-# nothing that was not already available.
-#
-# Guarded rather than asserted: xt_owner is a module, and a kernel without it
-# should leave a container that says why (the DNS self-check below), not one
-# that aborts here with an iptables usage error.
+# Guarded, not asserted: xt_owner is a module, and a kernel without it should
+# leave a container that says why (the DNS self-check below) rather than one
+# that aborts here.
 if iptables -A OUTPUT -p udp --dport 53 -m owner --uid-owner 0 -j ACCEPT 2>/dev/null; then
     iptables -A OUTPUT -p tcp --dport 53 -m owner --uid-owner 0 -j ACCEPT 2>/dev/null \
         || echo "WARNING: could not add the TCP/53 rule for root-owned sockets - large answers may fail" >&2

docs/ci.md

@@ -267,45 +267,39 @@ The rule underneath everything here:
 > **Every runner reachable from a repository that receives agent-authored
 > branches must be one you would accept an attacker running code on.**
 
-Note what that does *not* say. It says nothing about what the repository's
-workflows currently do, because a branch can add the first one. **A project
-with no CI at all is still exposed**, on a forge where its branches can reach
-a runner, and that is the case people do not expect, because there is nothing
-in the repository to look at.
+Note what that does *not* say: nothing about what the workflows currently do,
+because a branch can add the first one. **A project with no CI at all is still
+exposed** on a forge where its branches can reach a runner - the case people
+do not expect, because there is nothing in the repository to look at.
 
-So before pointing a sandbox at a project, ask one question: *what runners can
+So ask one question before pointing a sandbox at a project: *what runners can
 this repository's branches reach?*
 
-**If the answer is "hosted runners only"** (GitHub-hosted and nothing
-self-hosted attached), this is the mildest case, and it is worth saying so
-rather than alarming people. The runner is disposable and is not your machine,
-so "execution on a host you own" largely evaporates: there is no LAN behind
-it and nothing of yours on it. What remains is the third item, credentials:
-give pull-request runs no secrets they do not need and a read-only token
-(`permissions: contents: read`), and do not attach a self-hosted runner to
-that repository.
+**"Hosted runners only"** is the mildest case. The runner is disposable and is
+not your machine, so "execution on a host you own" largely evaporates - no LAN
+behind it, nothing of yours on it. What remains is credentials: give
+pull-request runs no secrets they do not need and a read-only token
+(`permissions: contents: read`), and attach no self-hosted runner.
 
-**If self-hosted runners are reachable**, that is the sharp case, and there are
-three ways to hold the rule, none of which require moving or renaming
-anything.
+**If self-hosted runners are reachable**, that is the sharp case. Three ways
+to hold the rule, none of which require moving or renaming anything.
 
-**1. Turn CI off for that repository, if it does not use CI.** Disabling the
-Actions unit means no workflow runs for that repository at all, so no runner is
-reachable and a branch cannot add the first workflow. Changing it back needs
-*admin* on the repository (Gitea's API puts repository settings behind
-`reqAdmin()`), so a bot account with Write cannot undo it. For a project that
-was not using CI anyway this costs nothing and closes the vector outright. It
-is the first thing to reach for and the most commonly overlooked.
+**1. Turn CI off for that repository, if it does not use CI.** No workflow
+runs at all, so no runner is reachable and a branch cannot add the first
+workflow. Turning it back on needs *admin* (Gitea puts repository settings
+behind `reqAdmin()`), so a bot account with Write cannot undo it. For a
+project not using CI this costs nothing, and it is the most commonly
+overlooked option.
 
 **2. Scope the dangerous runner down, rather than moving the repository out.**
 The instinct is to move agent-developed repositories somewhere with only
-hardened runners. For a public project that means breaking its name, its
-links and its ownership, which nobody should be asked to do.
+hardened runners, which for a public project breaks its name, links and
+ownership.
 
-Invert it. A runner can belong to an **owner** (a user as well as an
-organization) and then serves that owner's repositories; Gitea's runner model
-names the three cases: global (`OwnerID` and `RepoID` both zero), org/user
-level (`OwnerID` set), repository level (`RepoID` set). So:
+Invert it. A runner can belong to an **owner** - a user as well as an
+organization - and then serves that owner's repositories. Gitea's model has
+three cases: global (`OwnerID` and `RepoID` zero), org/user (`OwnerID` set),
+repository (`RepoID` set). So:
 
 - Make the *hardened* runner the owner-level default, so every repository gets
   it.
@@ -324,9 +318,9 @@ the vector is closed completely rather than bounded. The cost is one push
 after the review you were going to do anyway, and it is the only option here
 that does not depend on getting a forge's configuration right.
 
-If none of the three is available for a given project, that is worth knowing
-plainly: pointing a sandbox at it means accepting that a compromised session
-can run code on whatever that repository can reach.
+If none of the three is available for a given project, say so plainly:
+pointing a sandbox at it means accepting that a compromised session can run
+code on whatever that repository can reach.
 
 ## Egress
 
@@ -368,7 +362,7 @@ bridge subnet is one that check silently does not cover.
 
 With `runner.capacity: 1` one job occupies that network at a time. Above one,
 concurrent jobs share a broadcast domain and can reach each other's service
-containers, which is worth knowing before raising it.
+containers. Check that before raising it.
 
 The DNS and firewall shape is the same one dejima uses inside its own
 container (`docker/dns-proxy.sh` and `docker/init-firewall.sh`), applied one
@@ -499,10 +493,9 @@ What still applies:
   session commits locally and stops. The push grant *is* the CI
   code-execution grant; declining the first declines the second.
 
-Gitea 1.27's **scoped workflows** are worth knowing about here: a workflow
-maintained in a central repository the sandbox cannot push to, registerable
-instance-wide, so the workflow file itself sits outside any agent branch's
-reach. Use it, but it is not a gate. Scoped workflows run *alongside* a
+Gitea 1.27's **scoped workflows** apply here: a workflow maintained in a
+central repository the sandbox cannot push to, registerable instance-wide,
+so the workflow file itself sits outside any agent branch's reach. Use it, but it is not a gate. Scoped workflows run *alongside* a
 consuming repository's own workflows rather than replacing them, so a branch
 can still add a workflow of its own.
 

docs/configuration.md

@@ -18,7 +18,7 @@ what a given launch will actually use.
 
 | Key | Default | Purpose |
 | --- | --- | --- |
-| `[core] harness` | `claude-code` | agent harness the container runs |
+| `[core] harness` | *(none - must be set)* | agent harness the container runs; refused if unset, so a later default cannot change your selection |
 | `[core] forge` | `none` | forge integration: `gitea` (`gitea-pr`/`gitea-issue`), `github` (`gh`, opens `ssh.github.com:22` for push), or `none` (no forge at all) |
 | `[core] workflow` | `local-only` | workflow `sandbox-context.sh` tells the agent to follow: `worktree-pr` (push + open a PR), `branch-push` (push, stop at the branch), or `local-only` (commit locally, push nothing) |
 | `[core] command` | `claude` | container's main process (e.g. `nvim`), also settable per launch with `--cmd` |
@@ -27,6 +27,7 @@ what a given launch will actually use.
 | `[core] ssh_key` | unset | dedicated deploy key for headless/unattended git push |
 | `[core] known_hosts` | `~/.ssh/known_hosts` | known-hosts file used for git-over-ssh |
 | `[harness.claude-code] projects_dir` | `~/.claude/projects` | where Claude Code session transcripts live |
+| `[harness.claude-code] file_history_dir` | `~/.claude/file-history` | Claude Code's edit-undo record, mounted per session rather than whole |
 | `[harness.claude-code] statusline_path` | `~/.claude/statusline.sh` | statusline script path |
 | `[forge.gitea] ssh_host` | unset | Gitea/Forgejo SSH hostname: skip if you only push to GitHub |
 | `[forge.gitea] api` | unset | `https://<gitea host>/api/v1`: skip if you only push to GitHub |
@@ -39,6 +40,9 @@ what a given launch will actually use.
 | `[credentials] proxy` | `on` | host-side credential proxy; `off` mounts the real OAuth credential into the container directly, as before this feature |
 | `[credentials] transport` | `tls` | set to `loopback` to fall back to a plaintext forwarder with an explicit `ANTHROPIC_BASE_URL`, keeping the credential out of the container but reinstating the base-URL fingerprint: for when certificate pinning breaks the default TLS path (see [security notes](docs/security.md)) |
 | `[logging] loki_url` | unset | ship DNS/firewall logs to a Loki endpoint |
+| `[epitaph] enabled` | `on` | keep the summary Claude Code writes when a session compacts |
+| `[epitaph] dir` | `~/.local/share/dejima/epitaphs` | where they are kept; never mounted into a container |
+| `[epitaph] loki` | `off` | also push each epitaph to `[logging] loki_url`, host-side |
 | `[logging] conversation` | `off` | opt-in: also ship session transcripts to Loki (needs `loki_url`); see "Conversation logging (opt-in)" in docs/security.md |
 | `[audit] dir` | `~/.local/share/dejima/audit` | where exit-time audit findings are stored |
 | `[audit] baselines_dir` | `~/.local/share/dejima/baselines` | where audit baselines are stored |
@@ -93,32 +97,27 @@ name a release archive instead:
 dejima-dep add --global tarball:go@<sha256>:https://go.dev/dl/go1.25.0.linux-amd64.tar.gz
 ```
 
-To get the digest, fetch the archive on the host and hash it. Do not take a
-digest from the same page that serves the download unless you are willing to
-trust that page for both:
+Fetch the archive on the host and hash it. Do not take the digest from the
+same page that serves the download unless you trust that page for both:
 
 ```sh
 curl -fLO https://go.dev/dl/go1.25.0.linux-amd64.tar.gz
 sha256sum go1.25.0.linux-amd64.tar.gz
 ```
 
-Many projects publish a checksums file or sign their releases. Where they do,
-verifying against that is worth more than hashing whatever arrived, because it
-tells you the archive is the one the project published rather than the one
-your connection received.
+Where a project publishes a checksums file or signs its releases, verifying
+against that beats hashing whatever arrived: it says the archive is the one
+the project published, not the one your connection received.
 
 It unpacks with `--strip-components=1` into `/opt/dejima/<name>`, and
-`/opt/dejima/<name>/bin` goes on `PATH`. That one convention covers the usual
-shapes without the spec needing layout options: a Go archive has everything
-under `go/`, an nvim release under `nvim-linux-<arch>/`, and stripping one
-level puts `bin/` in the same place for both. Compression is detected from the
-archive, so `.tar.gz`, `.tar.xz` and `.tar.zst` all work.
+`/opt/dejima/<name>/bin` goes on `PATH` - one convention covering the usual
+shapes without layout options, since a Go archive is under `go/` and an nvim
+release under `nvim-linux-<arch>/`. Compression is detected, so `.tar.gz`,
+`.tar.xz` and `.tar.zst` all work.
 
-**The digest is mandatory**, and it is checked before anything is unpacked.
-Without one you would be approving a *name* whose contents the far end can
-change afterwards, which isn't a decision anyone can make once. With one, what
-you approved is a specific archive and nothing else can take its place. Two
-consequences worth knowing:
+**The digest is mandatory** and is checked before anything is unpacked.
+Without one you would approve a *name* whose contents the far end can change
+afterwards. Two consequences:
 
 - **A tarball dependency locks the project to one architecture**, because a
   digest does. A package name resolves per-host and a digest cannot: it names
@@ -128,7 +127,7 @@ consequences worth knowing:
   [#97](https://gitbox.branderson.io/dejima/dejima/issues/97).
 - **Upgrading is an approval, not a version bump.** A new release is a new
   digest, so it goes through `dejima-dep add` and `dejima-project approve`
-  again - which is the point, not friction to route around.
+  again. That is the mechanism working, not friction to route around.
 
 The URL's character set is deliberately narrower than URLs allow: the spec is
 interpolated into a `RUN` line that runs as root at build time, so anything
@@ -136,12 +135,11 @@ shell-active is refused outright rather than quoted around, and only `https`
 is accepted.
 
 **Declaring grants nothing.** The project directory is bind-mounted
-read-write, so a compromised session could write these files - they are read
-only to compare against reality and suggest a command. Approvals live
-outside the container, under `~/.config/dejima/` by default
-(`$DEJIMA_CONFIG_DIR` relocates the whole tree) - global at its
-root, per project under `projects/<slug>/` - and are written only by tools
-you run on the host:
+read-write, so a session could write these files; they are read only to
+compare against reality and suggest a command. Approvals live outside the
+container under `~/.config/dejima/` (`$DEJIMA_CONFIG_DIR` relocates the
+tree) - global at its root, per project under `projects/<slug>/` - written
+only by tools you run on the host:
 
 ```sh
 dejima-project status ~/myproject     # declared vs approved vs missing
@@ -329,6 +327,21 @@ nvim_config = ~/.config/nvim
 Set both in `~/.config/dejima/config` if you use this mode regularly, or pass
 `--cmd nvim` per launch once `nvim_config` is set.
 
+The lockfile is layered on as a **copy**, not as your file: `bin/dejima`
+mounts a `mktemp` duplicate and deletes it afterwards, so a plugin sync
+inside a session cannot rewrite the host's `lazy-lock.json`.
+
+`~/.local/share/nvim` (plugin source and `mason/` binaries), `~/.local/state/
+nvim` and `~/.cache/nvim` are **Docker volumes, not host directories**. Your
+own nvim data is never mounted, so nothing a session installs can reach the
+editor you run outside the container. They do persist across containers,
+which is what stops every launch from reinstalling plugins - so they are
+named per project. Under one shared name, a session that wrote a Lua file
+under `lazy/`, or replaced a binary under `mason/bin/`, would leave something
+the next session in an *unrelated* project executed. Nothing there is
+protection against a session tampering with its own project's plugins, which
+it can do freely and which only ever runs inside the sandbox.
+
 If your config references something outside that directory by absolute host
 path (a helper script in a dotfiles `bin/`, say), bring it in with
 `[mounts] readonly` (colon-separated, mounted read-only at the same path):
@@ -348,9 +361,9 @@ session is meant to write to and whose contents should outlive the container:
 read_write = ~/specs
 ```
 
-Both lists mount at the host path they name, which is the point: a convention
-that says "specs live in `~/specs`" keeps working inside the container without
-being rewritten. That is the difference from `[core] share_dir`, which is
+Both lists mount at the host path they name, so a convention that says
+"specs live in `~/specs`" keeps working inside the container without being
+rewritten. That is the difference from `[core] share_dir`, which is
 mounted at `~/shared` regardless of where it lives on the host because its job
 is to be a well-known handoff point.
 
@@ -366,6 +379,31 @@ resolving symlinks, so `read_write = ~` is caught too.
 Sessions are told which directories these are, so "everything else is
 disposable" stays true in the briefing rather than becoming a trap.
 
+### One-off mounts for a single launch
+
+`--extra-readonly PATH` and `--extra-read-write PATH` add a path for one run
+without editing the config. `--ro` and `--rw` are the short forms. Both are
+repeatable, one path per occurrence:
+
+```sh
+dejima --ro ~/other-project ~/myproject          # read while migrating code out
+dejima --rw ~/scratch --ro ~/reference ~/myproject
+```
+
+They **append** to the config's lists rather than replacing them, and they
+are the same lists, so everything above applies unchanged: the existence
+check, the refusal when a path overlaps something dejima keeps out of a
+session's reach, the refusal when one path appears in both lists, and the
+briefing sentence naming what a session can write. A flag says *where*, not
+*whether*, and there is no flag that waives a refusal.
+
+Two differences from the config form, both because a command line is not a
+file. A relative path is accepted and made absolute against the working
+directory, which a config file has no equivalent of. And a path containing a
+colon is refused outright: both lists are colon-separated, so such a path
+cannot be represented in one at all and would otherwise split into two
+mounts that each fail the existence check.
+
 ## Shared scratch space
 
 `~/shared` is a read-write bind mount of `[core] share_dir` (default
@@ -449,6 +487,14 @@ since the same session can be resumed unrestricted on the host via
 `claude --resume`. No-op outside the container. `deny-dangerous-bash.sh`
 (`PreToolUse`) also applies here, since `~/.claude/hooks` is mounted in.
 
+When GitHub is a push destination this session - `--allow-github-push`, or
+`forge = github` under a pushing workflow - the briefing also tells the
+session not to push branches or open pull requests on any project other than
+the one it was pointed at. Lifting the default gives an agent the ability
+without the reason it is normally off (see
+[docs/security.md](security.md#github-push-blocked-by-default-optional-opt-in)),
+so the reason is supplied where it will be read.
+
 `git-push-guard.sh` (`PreToolUse`/`Bash`, also `DEJIMA`-gated) blocks
 any `git push` whose resolved remote isn't the Gitea host, and any push to
 `main`/`master` even there. Like the other hooks, this is a fast-fail guard

docs/development.md

@@ -17,7 +17,7 @@ run them.
 | `docker/ulogd.conf` | `ulogd2` config: turns `init-firewall.sh`'s NFLOG-tagged packets into JSON |
 | `docker/log-shipper.sh` | Tails the dnsmasq/ulogd2 logs and ships them to the configured Loki endpoint (`[logging] loki_url`); also ships session transcripts when conversation logging (`[logging] conversation`) is on |
 | `docker/entrypoint.sh` | Runs as root: DNS proxy, firewall, logging, drops to the sandbox user via `gosu`, then execs the session's command |
-| `docker/claude-wrapper.sh` | Shadows `claude` on `PATH` to always add `--dangerously-skip-permissions` |
+| `docker/harness-wrapper.sh` | Shadows the selected harness's command on `PATH` to add its no-confirm flags and dejima's settings layer. Named by no harness; `entrypoint.sh` links it into place |
 | `docker/docker-compose.yml` | Wires up mounts/env; invoked via `bin/dejima`, not directly |
 | `hooks/sandbox-context.sh` | `SessionStart` hook: injects sandbox context/workflow reminders, and any `.dejima.md` the project ships |
 | `.dejima.md` | Project-level agent instructions, harness-agnostic: `CLAUDE.md`'s role under a neutral name. dejima loads a project's into the session context; this one is dejima's own |
@@ -34,6 +34,10 @@ run them.
 | `bin/dejima-notify-watch` | Host-side watcher: turns markers into ntfy pushes composed entirely from host data |
 | `bin/dejima-project` | Review and approve project domains and dependencies from the host |
 | `bin/dejima-project-lib.sh` | Shared library sourced by `bin/dejima-dep`, `bin/dejima-project`, `bin/dejima` and the remote-access scripts; not on PATH |
+| `bin/dejima-epitaph` | Host-side: keeps the summary a session leaves when its context is compacted away, in a store nothing mounts. Run from `bin/dejima`'s exit trap |
+| `bin/dejima-harness-lib.sh` | Loads the selected harness's plugin and provides the `emit_*` functions it declares through; the guarded-store refusal runs in `emit_mount`. Not on PATH |
+| `bin/harness/<name>.sh` | One per supported harness: what dejima does for that harness that no manifest value could state. Trusted computing base, sourced by `bin/dejima` |
+| `bin/dejima-slug-lib.sh` | The `~/.claude/projects` slug scheme and which slugs belong to a workspace; sourced by `bin/dejima` (what to mount) and `bin/dejima-epitaph` (whose transcripts to keep), so the two cannot disagree. Not on PATH |
 | `bin/dejima-ttyd` | Launches `ttyd` on loopback in front of `dejima-attach` |
 | `systemd/*.service` | User units for `dejima-ttyd` and `dejima-notify-watch`; ship disabled, installed by hand |
 | `docs/providers.md` | The manifest contract for the forge/workflow/harness axes, key by key, with the shipped manifests as worked examples |
@@ -95,7 +99,7 @@ Docker host: `runtime.bats` proves that a service bound to the host's
 loopback is unreachable from inside a container while one bound to the
 bridge gateway *is* reachable. The second half is a positive control - it is
 what stops the first from passing in an environment where the container
-simply reaches nothing. That pair is the ttyd front door's whole containment
+reaches nothing at all. That pair is the ttyd front door's whole containment
 argument, so run it before trusting a front door you just set up.
 
 Every mutation in `runtime.bats` (firewall rules, `ipset` entries) happens
@@ -110,6 +114,22 @@ socket withheld from the confined job, granted to the privileged job, and
 the egress each is allowed) - see [docs/ci.md](ci.md). It skips outright
 unless `DEJIMA_CI_TIER` is set, so it has nothing to say when run locally.
 
+### Branches
+
+Pull requests target `development`, which merges to `main` periodically, so
+`main` stays releasable while work lands quickly. `gitea-pr` defaults its
+base to the repository's default branch, so pass `--base development`.
+
+CI needs no change for this: `.gitea/workflows/tests.yml` triggers on
+`pull_request_target` with no branch filter, so it already runs whatever the
+base is.
+
+Nothing *enforces* the target. `git-push-guard.sh` blocks direct pushes to
+the branches a workflow names, and that list is workflow-wide rather than
+per-project - adding `development` would impose it on every dejima user,
+most of whom have no such branch. So this is a convention with a check
+attached: read what `gitea-pr` printed.
+
 ### Running the tests
 
 ```sh
@@ -130,3 +150,208 @@ push to one, `runtime.bats` included. That job needs two things from the
 runner that an ordinary one doesn't (the host's Docker socket and the host
 network namespace), and it checks for both up front rather than failing
 obscurely partway through; the file says why each is needed.
+
+## Adding a new harness
+
+`[core] harness` picks a directory under `harnesses/`, and a harness manifest
+declares four things. Everything else dejima does for Claude Code is written
+against Claude Code, and this is the inventory of it: what a second harness
+has to supply, and where dejima has to stop assuming. Taken from a pass over
+the tree, so each row names the file to change rather than describing a
+sketch.
+
+### What a harness must be able to do
+
+Support is per harness and deliberate. A supported harness gets the same
+depth of integration Claude Code has - the same mounts, the same guards, the
+same audit surfaces - so a harness is not a manifest someone drops in. That
+is only possible if the harness can host what those protections need.
+
+These are the requirements. A harness that fails one of the first five
+cannot be supported, and that is a decision about the harness rather than a
+gap in dejima.
+
+| | Requirement | Why it is load-bearing |
+| --- | --- | --- |
+| **R1** | **Unattended execution.** A documented mode that runs tool calls without confirming each one. | The premise the project rests on. A harness that stops to ask is one an operator sits in front of, and the argument for accepting that mode is that the container makes being unattended survivable. |
+| **R2** | **Tool-call interception.** A supported mechanism to inspect a command before it runs and refuse it, without patching the harness. | The push guard, the merge guard and the dangerous-command denylist all live here. They are guards against an eager agent rather than a boundary against a hostile one, but a harness without them ships three protections short and the operator cannot add them back. |
+| **R3** | **Host-controlled context at session start.** A way to place text into the session's context at start, from a file the operator controls and the session cannot rewrite for the run in progress. | How the briefing and the project's `.dejima.md` arrive. It is also how the agent learns it is sandboxed, which is what makes it ask for a domain instead of working around the firewall. |
+| **R4** | **State dejima can locate or discard.** Per-project state must be computable on the host before launch, so it can be scoped per project - or safely discardable, so it can stay container-local. | #104. A store that is neither computable nor discardable cannot be scoped, and mounting it whole reintroduces cross-project context poisoning. |
+| **R5** | **A withholdable credential.** The model credential must live somewhere dejima can decline to mount, and the harness must accept a substitute: a staged fake, or an environment variable pointed at the host-side proxy. | The credential never entering the container is a headline property, not a nicety. |
+
+Three more are required, but may be declared absent in the manifest, in
+which case dejima says at launch what is not in force:
+
+| | Requirement | What is lost without it |
+| --- | --- | --- |
+| **D1** | Session lifecycle signals - something that fires when the agent wants attention and when it stops. | Remote access notifications. The session still runs and is still attachable. |
+| **D2** | A parseable record of the session. | Epitaph, and conversation logging. |
+| **D3** | An enumerable config directory whose extension points can be mounted read-only. | Config parity, and the read-only mounts that stop a session installing a persistent hook for the next one. |
+
+Declaring one absent is a real decision with a visible cost, not a
+formality. It is there so a harness can be supported honestly at a smaller
+guard set rather than shipped with three protections quietly missing.
+
+**The guard rails are declared in the manifest and reported at launch.**
+`guards` lists which of `bash-deny`, `push-guard`, `merge-guard` and
+`notify` the harness can host. A name outside that set is refused, because a
+typo would otherwise turn a guard off in silence. The ones a harness does
+not declare are named once at launch, with what each costs - and nothing is
+printed at all when it hosts them all, so the message stays a signal rather
+than a line to scroll past.
+
+Core gates the work on the declaration rather than on whether a plugin
+happens to define the hook: a harness that forgot to declare a guard but
+implemented it would otherwise get a protection the launch message has just
+told the operator it does not have.
+
+### Worked example: why aider is not supported
+
+aider clears R1 (`--yes-always`), R3 (`--read` places a file in context),
+R4 (its state is `.aider.*` inside the workspace, which is already mounted)
+and R5 (it takes an API key from the environment).
+
+It fails **R2**. aider has no hook or event system; the only occurrence of
+"hook" in its options is `--no-verify`, which controls git's own pre-commit
+hooks. There is no supported point at which a command can be inspected and
+refused, so `git-push-guard.sh`, `merge-guard.sh` and
+`deny-dangerous-bash.sh` have nowhere to live and no equivalent to be
+rebuilt from. It also fails D1.
+
+The container would still bound an aider session. But three guards an
+operator gets everywhere else would be missing with no way to restore them,
+and "supported" would mean something different for that one harness than
+for every other. So it is not supported, and this is the reason.
+
+### Where a harness's support lives: three tiers
+
+| Tier | What | Rule |
+| --- | --- | --- |
+| **Core** | the mechanism, and every enforcement point | harness-agnostic, always |
+| **Manifest** | `harnesses/<name>/harness.conf` | data. Never sourced, never executed. Selects among named core mechanisms and supplies paths |
+| **Plugin** | `bin/harness/<name>.sh` | the harness-specific remainder, as code, in this repository |
+
+**A feature belongs in the manifest if it can be stated as a value.** It only
+moves to a plugin when it cannot. `projects_dir` and `file_history_dir` are
+both paths, so both are manifest keys; the rule that a *name* in the first
+selects an *entry* in the second is Claude Code's own and could not be
+written as a value, so that is the plugin. "Add a manifest key" is the
+answer until it demonstrably is not.
+
+**Plugins live under `bin/`, not beside the manifest.** `dejima-audit`'s
+`tcb_snapshot` covers `install.sh`, `bin/`, `hooks/`, `docker/` and `test/`,
+and not the provider directories. Host-side plugin code under `harnesses/`
+would be code that runs on the host and that a session could edit in
+dejima's own checkout without raising a CRITICAL - the exact class the audit
+exists to catch. Under `bin/` it is covered for free, and "manifests are
+data, never executed" stays true word for word.
+
+**Plugins are sourced, not invoked**, and are part of the trusted computing
+base. The risk worth designing against is not a careless plugin but three
+plugins implementing the security model three ways, so a plugin uses core's
+own primitives rather than re-deriving them. An invoked protocol would also
+put a new text parser on the launch path, and every parsing bug this
+repository has had is in that family.
+
+**The manifest may carry values that are only used as paths or compared as
+data. It must not carry text that becomes part of a parser's grammar, or the
+identity of a security surface.** `dejima-audit`'s `ack` recovers a finding's
+id by stripping a fixed label from the front of the finding text, anchored
+deliberately (#45). Those labels are grammar. Provider directories are not in
+`tcb_snapshot`'s list, so in the self-mount case a session can edit a
+manifest without raising a CRITICAL - which is why host-execution surfaces
+are a plugin hook (`harness_host_exec_surfaces`) rather than a manifest key.
+The label stays a literal in code the audit watches; it just lives with the
+harness that owns it.
+
+**A plugin computes; core enforces.** A plugin declares through `emit_mount`
+and `emit_env`, which validate at the point of declaration - `emit_mount`
+runs the guarded-store refusal there, so a harness state directory that
+overlapped the approval, audit or epitaph store is refused without each
+plugin having to remember. A plugin never touches `extra_mounts` itself, and
+a static test holds that. A plugin may *classify* an audit path; it may never
+suppress a finding.
+
+Every shipped harness has both a manifest and a plugin - a static test
+refuses one without the other, because a manifest alone would present as a
+supported harness while none of the work had been done.
+
+### What the manifest already covers
+
+`harnesses/<name>/harness.conf`, and the full contract is in
+[providers.md](providers.md):
+
+| Key | What it decides |
+| --- | --- |
+| `command` | the container's main process |
+| `keys` | this harness's own `[harness.<name>]` config schema |
+| `resume_command` | what the briefing tells a session to hand the operator |
+| `briefing` | this harness's fragment of the session briefing |
+
+### Reading a harness's own config
+
+`harness_config_get <key>` resolves whichever harness `[core] harness`
+names. Use it; do not spell a `[harness.<name>]` section out at a call site.
+
+`keys` is declared per harness, so a literal section name makes it a
+declaration only one harness can ever be read through - a second harness
+gets selected, gets validated, and is then handed the first one's paths with
+nothing erroring anywhere. Three call sites did exactly that, in
+`bin/dejima-audit` (twice) and `bin/dejima-epitaph`. A static test now holds
+the rule.
+
+It returns empty when no harness is selected. `bin/dejima` refuses to start
+in that state, so that is reachable only from the standalone tools, where a
+harness-specific path resolving to nothing is the honest answer.
+
+### The inventory
+
+| Area | What is Claude Code specific | Where |
+| --- | --- | --- |
+| Install | `npm install -g @anthropic-ai/claude-code`, and a pinned `CLAUDE_CODE_VERSION` build arg | `docker/Dockerfile` |
+| Permission bypass | the flag itself, now `no_confirm_args` in the manifest; the wrapper is generic | `harnesses/<name>/harness.conf` |
+| Settings layering | that overrides arrive as an extra layer rather than a mount over the operator's file (#47) is core; the flag that carries them is `settings_args` in the manifest | `harnesses/<name>/harness.conf` |
+| Telemetry | `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1` | `docker/Dockerfile` |
+| Egress allowlist | `platform.claude.com` always, `api.anthropic.com` only when no credential proxy is in front of it | `docker/fixed-domains.sh` |
+| Credentials | the whole proxy exists because Claude Code authenticates from `~/.claude/.credentials.json` rather than an env var; the entrypoint seeds a well-formed fake so it has something to read | `bin/dejima-anthropic-proxy`, `docker/anthropic-proxy-forwarder.js`, `docker/entrypoint.sh` |
+| Hook protocol | the event payload arrives on stdin; `SessionStart` returns `hookSpecificOutput.hookEventName`/`additionalContext` as JSON, while a `PreToolUse` hook blocks by exiting 2 with its reason on stderr | `hooks/*.sh` |
+| Hook registration | four hooks written into `~/.claude/settings.json` in Claude Code's schema, one `SessionStart` with a `startup\|resume\|clear\|compact` matcher and three `PreToolUse` matching `Bash`. Now in the plugin, gated per declared guard | `bin/harness/claude-code.sh` |
+| Briefing delivery | the briefing and the project's `.dejima.md` reach the session *through* the `SessionStart` hook. No hooks, no briefing | `hooks/sandbox-context.sh` |
+| Notifications | `Notification`/`Stop` hooks, registered per session into the settings layer, are what drop the markers remote access watches | `hooks/dejima-notify.sh`, `bin/dejima` |
+| Config parity | every `~/.claude` subdirectory is enumerated and mounted by name: `plugins`, `hooks`, `skills`, `commands`, `agents`, `output-styles` read-only; `backups`, `tasks`, `plans`, `sessions` read-write | `docker/docker-compose.yml` |
+| Staged config | `settings.json`, `CLAUDE.md` and `statusline.sh` are staged read-only under `/etc/dejima-claude` and installed into place | `docker/entrypoint.sh`, `bin/dejima` |
+| App state | `~/.claude.json` is copied rather than mounted, because it mixes live state with `mcpServers` config | `bin/dejima` |
+| State layout | the project key - cwd with every non-alphanumeric replaced by `-` - is Claude Code's scheme, and dejima reproduces it in four places | `bin/dejima-slug-lib.sh`, `bin/dejima-project-lib.sh`, `docker/log-shipper.sh` |
+| Per-project scoping | `projects/` is mounted one slug at a time and `file-history/` one session at a time, the second resolved by pairing `projects/<slug>/<id>.jsonl` with `file-history/<id>` (#104). Now in the plugin | `bin/harness/claude-code.sh` |
+| Transcript format | conversation logging transforms Claude Code's JSONL line by line | `docker/log-shipper.sh` |
+| Compaction format | epitaph keys on `isCompactSummary` records and the `compact_boundary` that precedes them | `bin/dejima-epitaph` |
+| Audit surfaces | `.claude/settings.json` is CRITICAL and the rest of `.claude/` is a WARNING, because that is where this harness keeps things that run | `bin/dejima-audit` |
+| Host code execution | `statusline.sh` is audited as its own CRITICAL surface because `settings.json`'s `statusLine.command` runs it **unsandboxed on the host**. Declared by the plugin, since a finding's label is a parser's grammar rather than data | `bin/harness/claude-code.sh` |
+
+### Two things worth deciding early
+
+**What runs on the host when the harness is configured.** `statusLine.command`
+is the one that exists today, and it is not a quirk: a harness that will
+execute an operator-configured command outside the container hands a session
+that writes that file a way onto the host. Find that surface in the new
+harness before supporting it, because it belongs in the audit from the first
+release rather than after someone finds it.
+
+**What the harness keeps, and who else can read it.** The per-project
+scoping of `projects/` and `file-history/` exists because a session that can
+write another project's memory can plant text a later session there reads as
+established background (#104). A new harness's state directory needs the same
+question asked of it, one directory at a time. Anything mounted host-wide
+should be there because it is keyed by session rather than by project, and
+because nothing loads it into a session's context at startup - which is the
+reasoning that leaves `tasks`, `plans`, `sessions` and `backups` mounted
+whole today.
+
+### Testing a new harness
+
+`providers.bats` covers the manifest contract itself and is where a new
+`harness.conf` gets its parsing and validation tests. The suites that assert
+on Claude Code paths by name - `config.bats`, `audit.bats`, `static.bats`,
+`runtime.bats` - are the ones that will need a second set, and a green run
+against the existing harness proves nothing about a new one. Nothing in
+`test/` currently runs a second harness end to end.

docs/providers.md

@@ -211,7 +211,11 @@ A workflow's `briefing.md` is plain Markdown plus two marker forms that
 
 | Key | Type | Default | Meaning |
 | --- | --- | --- | --- |
-| `command` | `identifier` | `claude` | The container's main process for this harness. |
+| `command` | `identifier` | `claude` | The container's main process for this harness, and the name the wrapper shadows on `PATH`. `[core] command` defaults to it. |
+| `command_path` | `?path` | (empty) | Where the real binary lives, behind the shadow. |
+| `guards` | `?list:identifier` | (empty) | Which of dejima's guard rails this harness can host: `bash-deny`, `push-guard`, `merge-guard`, `notify`. A name outside that set is refused rather than treated as an unknown guard, since a typo would otherwise turn a guard off silently. The ones a harness does not declare are named at launch - see "What a harness must be able to do" in [development.md](development.md). |
+| `no_confirm_args` | `?template` | (empty) | The flags that make the harness run without confirming each tool call. Required in practice: dejima refuses to launch a harness that names none, because unattended operation is the premise the container is built on (R1 in [development.md](development.md)). |
+| `settings_args` | `?template` | (empty) | How this harness is told to read dejima's own settings layer, mounted at `/etc/dejima-settings.json`. Empty for a harness that takes its overrides some other way. |
 | `keys` | `?list:keyspec` | (empty) | This harness's own `[harness.<name>]` config schema - same grammar as a forge's `keys`. |
 | `resume_command` | `?template` | (empty) | Prose template (may use `{{session_id}}`) shown at the end of the briefing for resuming this exact session on the host. |
 | `briefing` | `?relpath` | (empty) | This harness's briefing fragment file, relative to its own directory. |
@@ -219,14 +223,22 @@ A workflow's `briefing.md` is plain Markdown plus two marker forms that
 Worked example, `harnesses/claude-code/harness.conf`:
 
 ```
-command        = claude
-keys           = projects_dir:path=~/.claude/projects statusline_path:path=~/.claude/statusline.sh
-resume_command = claude --resume {{session_id}}
-briefing       = briefing.md
+command         = claude
+command_path    = /usr/local/share/npm-global/bin/claude
+guards          = bash-deny push-guard merge-guard notify
+no_confirm_args = --dangerously-skip-permissions
+settings_args   = --settings /etc/dejima-settings.json
+keys            = projects_dir:path=~/.claude/projects file_history_dir:path=~/.claude/file-history statusline_path:path=~/.claude/statusline.sh
+resume_command  = claude --resume {{session_id}}
+briefing        = briefing.md
 ```
 
-Reading it: the container's main process is `claude`; `[harness.claude-code]`
-gets `projects_dir` and `statusline_path`, both `path`-typed with defaults;
+Reading it: the container's main process is `claude`, wrapped on `PATH` in
+front of the real binary at `command_path` so that every invocation - not
+only the one the entrypoint makes - gets `--dangerously-skip-permissions`
+and dejima's settings layer; `[harness.claude-code]`
+gets `projects_dir`, `file_history_dir` and `statusline_path`, all `path`-typed
+with defaults;
 resuming the session on the host is `claude --resume <session id>`, with
 `{{session_id}}` filled in at composition time; and its briefing fragment is
 `briefing.md`.
@@ -318,30 +330,21 @@ every launch:
   that has config keys but none of them resolve to a host is refused with
   "configure it, or select `forge = github`, or `forge = none` with a
   workflow that does not push".
-- A forge whose `requires` keys are **partly** set - some configured, some
-  not (a key written but left empty, as `install.sh`'s commented-out starter
-  config makes easy, counts as not set: it resolves to nothing, and every
-  call the forge makes with it would fail inside the container) - is refused
-  as a likely typo, but only when the workflow actually
-  pushes: a non-pushing workflow never invokes the forge, so e.g. gitea's
-  `ssh_host` alone (set only so the firewall allows fetches) is a legitimate
-  configuration, not a typo. A wholly unconfigured forge is never flagged by
-  this rule - that's the case above.
-- A pushing workflow whose forge has **none** of its required settings
-  configured is refused too, even when `fixed_hosts` means the "resolves to
-  no host" rule above never fires. This is what catches `github`: its single
-  `requires` key (`push_token_file`) means "partly configured" (the rule
-  above) can never trigger for it either - there's nothing to be partial
-  between - so without this rule a pushing workflow could select `github`
-  with no `push_token_file` set and launch anyway, `fixed_hosts` masking the
-  fact that nothing was actually configured.
-
-Two qualifications worth stating explicitly for the "partly configured" rule:
-it only applies when the workflow's `push = on` (a non-pushing workflow never
-touches the forge, so an incomplete `[forge.<name>]` section is harmless),
-and a required key with a non-empty schema default - gitea's `token_file`,
-defaulting to `~/.config/gitea/token` - never counts as missing, since its
-resolved value is already correct even when the operator never wrote it.
+- A forge whose `requires` keys are **partly** set is refused as a likely
+  typo, but only when the workflow pushes. A key written but left empty
+  counts as not set - it resolves to nothing, and every call made with it
+  would fail inside the container. A non-pushing workflow never invokes the
+  forge, so gitea's `ssh_host` alone (set only so the firewall allows
+  fetches) is legitimate. A required key with a non-empty schema default -
+  gitea's `token_file` - never counts as missing, since its resolved value is
+  correct even unwritten. A wholly unconfigured forge is the case above, not
+  this one.
+- A pushing workflow whose forge has **none** of its required settings is
+  refused too, even where `fixed_hosts` stops the "resolves to no host" rule
+  firing. This catches `github`: with a single `requires` key there is
+  nothing to be partial between, so without this rule a pushing workflow
+  could select it with no `push_token_file` and launch anyway, `fixed_hosts`
+  masking that nothing was configured.
 
 ## Shipped providers
 
@@ -353,7 +356,7 @@ resolved value is already correct even when the operator never wrote it.
 | workflow | `worktree-pr` | worktree → push → PR, stops at the PR |
 | workflow | `branch-push` | worktree → push, stops at the branch |
 | workflow | `local-only` | **default** - commits locally, pushes nothing |
-| harness | `claude-code` | **default**, and the only harness today |
+| harness | `claude-code` | the only harness today, and **not a default** - `[core] harness` must be set |
 
 **The shipped defaults are deliberately the non-pushing pair**
 (`forge = none`, `workflow = local-only`). A fresh install has no forge host

docs/remote-access.md

@@ -44,7 +44,7 @@ What it does **not** add is a way *out* of a sandbox. Configured as
 described below, a compromised session cannot reach `ttyd`, cannot reach
 your proxy, and gains nothing from the fact that your phone can: it has a
 route to exactly one host address, and neither front door is on it. That
-property is worth stating precisely because it is the one a plausible-looking
+property needs stating exactly, because it is the one a plausible-looking
 mistake inverts - a proxy bound to `0.0.0.0` or to a bridge address is a
 front door opening *inward*, and that is a full sandbox escape rather than a
 degradation. [Verifying it](#verifying-it) is two commands, and worth
@@ -304,7 +304,7 @@ also what catches a distro `ttyd.service` still running on that port, which
 is why this is worth doing even if you are confident about the proxy. `test/runtime.bats`
 asserts this same containment property against a live container (with a
 positive control on the same path, so a passing result can't come from the
-container simply reaching nothing), and it passes on a real Docker host.
+container reaching nothing), and it passes on a real Docker host.
 
 ## SSH front door
 
@@ -700,16 +700,14 @@ what the outer proxy sends.
 
 ## Notifications
 
-`dejima` registers `hooks/dejima-notify.sh` as a `Notification` and
-`Stop` hook for a remote-enabled session - into the throwaway settings
-override layer it already generates for that session, never into your real
-`~/.claude/settings.json`, which every sandbox and your host's own Claude
-Code read directly. Your own `Notification` and `Stop` hooks keep running
-alongside it: Claude Code merges hooks across settings layers rather than
-letting the higher one replace the lower. That layer is mounted read-only, so a
-session cannot unregister its own hook. The hook writes one empty file, named `waiting` or
-`done`, into this run's own bind-mounted marker directory. No content, no
-stdin, no network.
+`dejima` registers `hooks/dejima-notify.sh` as a `Notification` and `Stop`
+hook for a remote-enabled session, into the throwaway settings override layer
+it already generates - never into your real `~/.claude/settings.json`. Your
+own hooks keep running alongside it, since Claude Code merges hooks across
+layers rather than replacing. The layer is mounted read-only, so a session
+cannot unregister its own hook. The hook writes one empty file named
+`waiting` or `done` into this run's marker directory: no content, no stdin,
+no network.
 
 On the host, `dejima-notify-watch` turns those markers into ntfy pushes.
 Same installation shape as above:
@@ -726,14 +724,13 @@ loop plus a five-second sleep). Repeat markers for the same session and
 event inside 60 seconds are debounced (`[remote-access] debounce` overrides
 that).
 
-One environment gotcha, because it fails silently: the watcher and
-`dejima` both resolve the marker root as `${XDG_RUNTIME_DIR:-/tmp}/dejima`,
-so they only meet when they agree on `XDG_RUNTIME_DIR`. systemd sets it for
-user units and a login shell gets it from pam_systemd, but a `dejima`
-run from a context that has neither - a bare `su`, a cron job, a tmux server
-started before login - writes its markers under `/tmp/dejima` while the unit
-watches `/run/user/<uid>/dejima`, and no push is ever sent, with nothing
-logged to say so. `systemctl --user show-environment` is where to check.
+One gotcha, because it fails silently: both sides resolve the marker root as
+`${XDG_RUNTIME_DIR:-/tmp}/dejima`, so they only meet when they agree on
+`XDG_RUNTIME_DIR`. systemd sets it for user units and a login shell gets it
+from pam_systemd, but a run from a context with neither - a bare `su`, cron,
+a tmux server started before login - writes to `/tmp/dejima` while the unit
+watches `/run/user/<uid>/dejima`. No push is sent and nothing is logged.
+Check with `systemctl --user show-environment`.
 
 The same directory is why you probably want lingering on:
 
@@ -741,23 +738,19 @@ The same directory is why you probably want lingering on:
 loginctl enable-linger "$USER"
 ```
 
-Both units here are *user* units. Without lingering they are stopped at
-logout and `/run/user/<uid>` is destroyed with the session, taking the
-session registry and every marker with it - so a sandbox left running in a
-detached tmux session has no front door and sends no notification until you
-log in again. It fails closed rather than open (nothing is left listening
-unguarded), but it fails invisibly, and it is the same `XDG_RUNTIME_DIR`
-that the gotcha above is about.
+Both are *user* units. Without lingering they stop at logout and
+`/run/user/<uid>` is destroyed with the session, taking the registry and
+every marker with it - so a sandbox left in a detached tmux session has no
+front door and sends nothing until you log in again. It fails closed, but
+invisibly, and on the same `XDG_RUNTIME_DIR` as the gotcha above.
 
 **The marker directory is a write channel into your runtime tmpfs**, and
-nothing bounds it. `/run/user/<uid>` is RAM-backed and sized at a fraction
-of physical memory, and each session gets its own directory under it
-bind-mounted read-write, so a session that writes into `/run/dejima-notify`
-until it is full degrades your whole login session - D-Bus, the systemd user
-manager - and not just dejima. A session can already fill the workspace
-mount, so this is an increment rather than a new kind of exposure, but a
-RAM-backed filesystem fails harder and sooner than a disk. There is
-deliberately no quota here; `rm -rf` on that run's marker directory (or
+nothing bounds it. `/run/user/<uid>` is RAM-backed and sized at a fraction of
+physical memory, so a session filling `/run/dejima-notify` degrades your
+whole login session - D-Bus, the systemd user manager - not just dejima. A
+session can already fill the workspace mount, so this is an increment rather
+than a new exposure, but a RAM-backed filesystem fails harder and sooner.
+There is deliberately no quota here; `rm -rf` on that run's marker directory (or
 logging out, which discards the tmpfs) is the recovery.
 
 **What a notification contains.** A title of `dejima: <project> needs input`
@@ -799,10 +792,11 @@ the next one is worth checking.
 2. **Is the hook registered?** The `Notification` and `Stop` entries live in
    the throwaway override layer, not in your real `settings.json`:
    `grep -c dejima-notify /etc/dejima-settings.json` from inside should be 2.
-   That file is passed to Claude Code as `--settings` by `claude-wrapper.sh`;
+   That file is passed to the harness by `harness-wrapper.sh`, as whatever
+   its manifest's `settings_args` names (`--settings` for Claude Code);
    if it is missing entirely, this container was not started by `dejima`.
-3. **Does the marker reach the host?** This is the one worth knowing, because
-   it splits the chain cleanly in half. From inside the sandbox:
+3. **Does the marker reach the host?** This is the one to check first: it
+   splits the chain cleanly in half. From inside the sandbox:
 
    ```sh
    : > "$DEJIMA_NOTIFY_DIR/waiting" && sleep 2 && ls "$DEJIMA_NOTIFY_DIR"

docs/responsible-use.md

@@ -57,10 +57,9 @@ at all.
 That second one is not a judgement about how good the containment is, and it
 does not get better with a version bump. There is no containment that fully
 prevents exfiltration: a determined attacker with code execution and any
-outbound path will find a way, and the allowlist narrows the paths rather
-than closing them. So the question is never whether you trust dejima enough.
-It is whether you can accept the risk of the contents leaving. If you cannot,
-do not hand them to an agent that could send them.
+outbound path will find a way. So the question is not whether you trust
+dejima enough, but whether you can accept the risk of the contents leaving.
+If you cannot, do not hand them to an agent that could send them.
 
 ## Look at what else is in the directory.
 
@@ -84,11 +83,9 @@ directory not to hand over.
 ## An allowlist entry is a grant, not a convenience.
 
 For every allowlist entry you give a dejima agent, consider what an attacker
-could do with it. That is the whole rule. What follows is why the answer
-varies so much between entries that look similar.
-
-The variable is what happens if the call is wrong, whether wrong means a
-prompt-injected instruction or an ordinary agent mistake.
+could do with it. The answer varies enormously between entries that look
+similar, and what it turns on is what happens if the call is wrong, whether
+wrong means a prompt-injected instruction or an ordinary agent mistake.
 
 A package registry or a docs site can be wrong without being expensive. A
 session that fetches the wrong package version, or scrapes a stale or
@@ -101,10 +98,9 @@ different shape entirely. A single call against any of them is the action,
 not a step toward one you'll review later: the deployment already ran, the
 row is already written or gone, the charge already happened. There is no
 diff downstream that catches it, because nothing about it passes through
-your repository at all. Allowlisting one of these isn't giving the session a
-faster way to fetch data; it's giving an unattended process standing
-authority to act on the world, unsupervised, the same authority you'd think
-twice about handing an unsupervised person.
+your repository at all. Allowlisting one of these gives an unattended process
+standing authority to act on the world, the same authority you would think
+twice about handing a person.
 
 Ask, before adding an entry: if this call happened by mistake, is the result
 something I fix in the next commit, or something I have to explain to
@@ -125,15 +121,14 @@ some vaguer sense. In this repository, dejima-ci-hardener resolves the
 CI-side egress
 allowlist by reading `.dejima/allowed-domains` from `origin/main`, never
 from a branch or a local checkout, and treats whatever it finds there as
-already reviewed. That trust isn't a policy statement, it's load-bearing:
-the only way content lands on `main` is through a pull request you merged,
-so reading from `main` is how the tool avoids asking a branch what it's
-allowed to do. Merging something onto `main` doesn't just ship it to users.
-It feeds a trust anchor that other tooling, this one and potentially future
-ones, is entitled to rely on without re-checking. Keeping insecure or broken
-code off `main` is what makes that reliance justified in the first place,
-and that job is yours; nothing about running the agent in a sandbox does it
-for you.
+already reviewed. That trust is load-bearing: the only way content lands on
+`main` is through a pull request you merged, so reading from `main` is how
+the tool avoids asking a branch what it's allowed to do. So merging onto
+`main` ships it to users and feeds a trust anchor that other tooling, this
+one and potentially future ones, is entitled to rely on without re-checking.
+Keeping insecure or broken code off `main` is what makes that reliance
+justified, and that job is yours; nothing about running the agent in a
+sandbox does it for you.
 
 ## Give it its own identity, then own what you merge.
 
@@ -150,10 +145,10 @@ actually holds a session to a smaller set of permissions than yours.
 
 Once the identity is separate, every commit and pull request the session
 produces is visibly labeled as the agent's own work, not yours, in your
-project's history. That visibility doesn't shrink your responsibility for
-it; it just makes it legible. Anyone reading your `main` later sees your
-name on the merge commit regardless of whose name is on the authored one,
-and will hold you to it accordingly.
+project's history. That visibility makes your responsibility legible without
+shrinking it. Anyone reading your `main` later sees your name on the merge
+commit regardless of whose name is on the authored one, and will hold you to
+it accordingly.
 
 ## Understand what you ship.
 
@@ -162,27 +157,27 @@ cheap enough that reading every diff starts to feel like the slow part of
 the loop, and skipping it works fine, repeatedly, until the one time it
 doesn't. A reviewer who reads every change catches most of what's wrong with
 it. A reviewer who only reads the ones that felt risky catches none of the
-ones that didn't feel risky, which is precisely the shape a subtle bug, or a
+ones that didn't feel risky, which is the shape a subtle bug, or a
 prompt-injected change dressed up as something ordinary, takes. Nothing
 about a faster loop changes what engineering judgment is for.
 
 ## This is not a blastproof box.
 
-The useful way to read this rule is as a statement about what kind of
-failure to plan for, not as a disclaimer to skim past. dejima's containment
-is Linux namespaces and a small capability set around the session, the same
-kind of isolation a container gives any process, not a virtual machine
-boundary; it shares the host kernel with everything else on the machine.
+Read this rule as a statement about what kind of failure to plan for.
+dejima's containment is Linux namespaces and a small capability set around
+the session, the same kind of isolation a container gives any process, not a
+virtual machine boundary; it shares the host kernel with everything else on
+the machine.
 That buys real things (see [docs/security.md](security.md) for the specific
 list: no working `sudo`, a deny-by-default firewall, read-only mounts over
 the surfaces that matter most), and every one of them assumes the container
 itself stays intact. A determined attacker, or a single container-escape
 vulnerability, doesn't erode that protection gradually; it removes every
 layer described in that document at once, because they all sit on the same
-kernel. The operating assumption to hold isn't "the walls will probably
-hold." It's "if this container is ever compromised at the kernel level,
-treat everything reachable from this host account as compromised with it,"
-which is a very different thing to plan around than an ordinary bug.
+kernel. So the assumption to hold is not that the walls will probably hold.
+It is that if this container is ever compromised at the kernel level,
+everything reachable from this host account should be treated as compromised
+with it, which is a very different thing to plan around than an ordinary bug.
 
 ## dejima's security says nothing about your software's.
 
@@ -195,12 +190,12 @@ SQL injection, hand-roll a broken authentication check, hardcode a secret,
 or pull in a compromised dependency, because those are properties of the
 code itself, decided at the moment it's written, not something a sandbox
 sitting around the process that wrote it can retrofit afterward. If what got
-built inside dejima turns out to be insecure, that isn't a containment
-failure; containment was never the control that was supposed to catch it.
+built inside dejima turns out to be insecure, containment was never the
+control that was supposed to catch it.
 
-**It is not something a future agent can retrofit either**, and that is the
-more useful half of this. "We will do a security pass later" was already a
-weak plan before agents made it cheap to say. Early architectural decisions
+**It is not something a future agent can retrofit either.** "We will do a
+security pass later" was already a weak plan before agents made it cheap to
+say. Early architectural decisions
 close off the fixes for whole classes of problem: where trust boundaries sit,
 what is authenticated and where, what data crosses which process, what a
 single compromised component reaches. A later pass can find a missing bounds
@@ -217,5 +212,4 @@ the one thing here no tooling in this repository is aiming at.
 
 The same practices you'd apply to code from any other source, review,
 linting, dependency scanning, tests aimed at the failure modes you care
-about, still have to happen here. dejima doesn't stand in for any of them;
-it operates on a different axis entirely.
+about, still have to happen here. dejima does not stand in for any of them.

docs/security.md

@@ -10,6 +10,12 @@ at repositories you already trust. Everything below is about how bad a
 compromised session *inside* that trust boundary can get, not a claim that
 one can't happen.
 
+**And none of it has been reviewed by anyone else.** This is early
+development: the reasoning here is one person's, the tests assert what that
+person thought to check, and no outside audit or adversarial testing has
+happened. Read what follows as claims awaiting scrutiny rather than claims
+that have survived it. See [SECURITY.md](../SECURITY.md).
+
 ## CI: push access is code execution elsewhere
 
 The controls in this document bound what a session can reach *directly*. They
@@ -113,9 +119,9 @@ table goes back to being decorative. `write:repository` and `write:issue` are
 the scopes its token needs (README, "Gitea/Forgejo PR workflow"), and nothing
 about that changes for a collaborator.
 
-One consequence is worth knowing before you meet it, because it presents as a
-misconfiguration and the obvious fix is the wrong one. Parts of Gitea's API
-are gated on *ownership* rather than on permission: `GET
+One consequence presents as a misconfiguration, and the obvious fix is the
+wrong one. Parts of Gitea's API are gated on *ownership* rather than on
+permission: `GET
 /repos/{owner}/{repo}/actions/runs` answers a non-owner with `"user should be
 the owner of the repo"` no matter which scopes the token carries. Read that
 way it looks like a scope error, and the reflex is to widen the token. That
@@ -212,27 +218,34 @@ hands to whichever session starts against it. See
   appears in `.git/config` directly), `!`-shell aliases - plus content
   changes under the workspace's `.claude/`, `.mcp.json`, `.envrc`,
   `.vscode/`, and `.git/hooks/` (symlinks included - a repointed symlink is
-  hashed by its target, not just its content), plus cross-project writes
-  under `~/.claude/projects` (see below). When the mounted workspace *is*
-  this repo's own sandbox itself (see the trusted-computing-base note
-  further below), it also diffs content hashes of the sandbox's own trusted
-  computing base - `bin/`, `hooks/`, `docker/`, `install.sh`,
-  `test/` - under that workspace. It's **advisory only**: it fires *after*
-  the session has already ended, so a report of a planted `core.hooksPath`
-  or hook file is something to read and clean up *before* the next `git`
-  or `claude` in that repo, not something that was blocked. It's silent when nothing
-  changed, and never affects the session's own exit code even if it can't
-  complete. It doesn't cover `[core] share_dir` (see
-  [Shared scratch space](configuration.md#shared-scratch-space)) - that
-  hand-off path stays unaudited - and its coverage is a
-  fixed list of known Claude Code config surfaces, not a general integrity
-  check, so a future config surface this list hasn't been taught about yet
-  would go unreported. **Its diff is time-windowed, not session-attributed:**
-  a finding means one of these paths changed while the session was running,
-  not that the session made the change. A concurrent dejima or Claude Code
-  session, another host process, or you editing the same files by hand
-  during the run all land in the same diff and get reported identically -
-  the tool has no way to tell them apart.
+  hashed by its target, not just its content). When the mounted workspace is
+  this repo itself, it also hashes the sandbox's own trusted computing base -
+  `bin/`, `hooks/`, `docker/`, `install.sh`, `test/`.
+
+  Worth knowing about that first surface, because the sandbox's own workflow
+  makes it easy to get wrong: **a git worktree isolates the working tree, not
+  `.git/config`.** Config is shared with the parent checkout, so a bare
+  `git config <key> <value>` inside a worktree writes the real repository's
+  config, including any exec key above. The briefing now says so, and gives
+  `git -C <path> config`. A guard against an unaware agent, not a boundary:
+  `.git/config` is in the read-write project mount and cannot be read-only
+  without breaking `git push -u`, which writes `branch.<name>.merge` on every
+  push. Detection covers the rest, and did - this was found by the audit
+  reporting a `.git/config` change.
+
+  **Advisory only.** It fires after the session ended, so a planted
+  `core.hooksPath` is something to clean up before the next `git` or
+  `claude` in that repo, not something that was blocked. Silent when nothing
+  changed, and never affects the session's exit code. It does not cover
+  `[core] share_dir` (see
+  [Shared scratch space](configuration.md#shared-scratch-space)), and its
+  coverage is a fixed list rather than a general integrity check, so a config
+  surface it has not been taught about goes unreported.
+
+  **Its diff is time-windowed, not session-attributed:** a finding means a
+  path changed while the session ran, not that the session changed it. A
+  concurrent session, another host process, or your own editing all land in
+  the same diff, reported identically.
 - **Approving and ignoring are different decisions, and the difference is
   what you are promising about next time.** Approving records the exact
   content: `safe <id> <sha>`, so this file with these bytes is fine and any
@@ -246,18 +259,22 @@ hands to whichever session starts against it. See
   [a] approve this exact content  [i] ignore this key from now on  [s] skip (ask again next time)  [q] quit
   ```
 
-  Skipping is the third option on purpose: a finding you have not decided
-  about is asked again next session, which is the right default for anything
-  you are unsure of. A WARNING-tier git config change is curatable the same
-  way, which matters for keys some tools rewrite on every run: without a
-  decision recorded against one, it re-fires every session and pushes real
-  findings out of the capped list.
+  A finding whose file has since been deleted is **cleared** rather than
+  approved: there is no content left to pin an approval to, and a surface
+  that does not exist does not run. Nothing is recorded, so if it comes back
+  it comes back as a new finding. Refusing instead - the old behaviour - left
+  the report unresolvable, and because `ack` opens CRITICAL reports first it
+  hid every other report behind that one.
 
-  Severity is not negotiable by either verb: an exec-capable git config key or
-  a shell alias is classified CRITICAL whatever the baseline says, and an
-  `ignore` entry is refused against a CRITICAL surface when the baseline is
-  read, not just when it is written. So neither decision can quietly downgrade
-  something that matters.
+  Skipping asks again next session, the right default when you are unsure.
+  It matters for keys some tools rewrite on every run: with no decision
+  recorded, one re-fires every session and pushes real findings out of the
+  capped list.
+
+  Severity is not negotiable by either verb. An exec-capable key or shell
+  alias is CRITICAL whatever the baseline says, and an `ignore` entry is
+  refused against a CRITICAL surface when the baseline is read, not only when
+  written - so neither decision can quietly downgrade something that matters.
 
 - `dejima-audit ack` shows what it is asking you to approve. For a surface
   that changed during the session it prints a unified diff of the
@@ -271,21 +288,19 @@ hands to whichever session starts against it. See
   live on disk); a removed one stores nothing at all, because a deletion
   diff *is* the whole former file.
 
-  Both the staging directory and the retained diffs live under
-  `$AUDIT_DIR` (`~/.local/share/dejima/audit` by default), never under
-  `/tmp`. That is the boundary the whole scheme rests on:
-  `docker-compose.yml` mounts nothing covering it, so the container cannot
-  read the retained content of earlier runs - `test/static.bats` pins that
-  as a regression guard. If the store is configured *inside* the mounted
-  workspace it becomes readable and deletable from within the sandbox, so
-  in that case no content is retained at all - only a note saying why. A
-  session killed with `SIGKILL` (no EXIT trap runs, so no `compare`)
-  leaves its staging directory behind; the next run against that
-  workspace reaps it, identifying dead owners by pid and process start
-  time recorded in an `.owner` file, so a concurrent session against the
-  same workspace is never mistaken for dead and disturbed. A directory
-  whose owner record is missing or unreadable falls back to a 24-hour age
-  check instead of guessing liveness.
+  Staging and retained diffs live under `$AUDIT_DIR`
+  (`~/.local/share/dejima/audit`), never `/tmp`. That is the boundary the
+  scheme rests on: nothing in `docker-compose.yml` covers it, so the
+  container cannot read earlier runs' retained content, and `static.bats`
+  pins that. A store configured *inside* the workspace is readable and
+  deletable from the sandbox, so nothing is retained in that case - only a
+  note saying why.
+
+  A `SIGKILL`ed session runs no EXIT trap and leaves its staging directory;
+  the next run against that workspace reaps it, identifying dead owners by
+  pid and process start time from an `.owner` file so a concurrent session is
+  never mistaken for dead. A missing or unreadable owner record falls back to
+  a 24-hour age check rather than guessing.
 
   Content is withheld from the store in several further cases, and in
   every one `ack` prints the reason in place of the diff - so an absent
@@ -300,101 +315,121 @@ hands to whichever session starts against it. See
   existed before the session but was too large to have been staged (over
   1 MiB) withholds with "no pre-session copy", not a false claim that it
   is new. The *current* side is bounded by that same 1 MiB cap, and - like
-  the staged side - the size is checked *before* any copy is taken: a file
-  the session grew past it withholds with "change too large to retain"
-  rather than being copied through `$TMPDIR` first. That ordering is the
-  point, not the cap: `compare` runs from the exit trap on every session
-  exit and the size is the session's own choice, so copying first would
-  let a session turn "I filled my own workspace" into that many bytes
-  written to `$TMPDIR` - frequently a tmpfs, i.e. host RAM - at exit, and
-  into a whole copy of the file left sitting there if that exit was a
-  `SIGKILL`. A generated diff is withheld if it is binary, over 200 lines
+  the staged side - checked *before* any copy is taken: a file the session
+  grew past it withholds with "change too large to retain" rather than
+  being copied through `$TMPDIR` first. The ordering is the point, not the
+  cap. `compare` runs from the exit trap on every exit and the size is the
+  session's own choice, so copying first would turn "I filled my own
+  workspace" into that many bytes written to `$TMPDIR` - often a tmpfs, so
+  host RAM - and into a whole copy left there if the exit was a `SIGKILL`.
+  A generated diff is withheld if it is binary, over 200 lines
   or 64 KiB (and is bounded to that as it is generated, for the same
   reason), or matches a best-effort credential screen - vendor prefixes
   including `sk-ant-`, `sk-proj-`, `ghp_`, `gho_`, `github_pat_`,
   `glpat-`, Slack's `xox[baprs]-`, `AIza`, `AKIA`, `_authToken`,
   `Authorization: Bearer`, and PEM private-key headers, matched
-  case-insensitively (deliberately: the one high-value surface this is
-  the only net for, `.claude/settings.json`'s `env` block, conventionally
-  holds its keys upper-case). That screen is still explicitly best-effort
-  and has real false negatives - the denylist above is the reliable half
-  of the defence, this is the second net. A git config key holding more
-  than one value on either side of the session is not compared at all:
-  git resolves such a key to its *last* value in *file* order, and the
-  audit does not resolve it - it parses the config files and sorts, so
-  lexicographically-last is not the value git will run, and the
-  pre-session side is every value the key had rather than one. Comparing
-  anyway produced a diff over values git does not use: with two
-  pre-session values it read as a deletion that never happened, while the
-  value the session actually appended - the one git now executes -
-  appeared nowhere in it. That case withholds with "this key has multiple
-  values and git uses the last", and must be compared by hand;
-  `include.path` is legitimately multi-valued, so this is an ordinary
-  shape, not only a planted one. A generated diff can also still come out
-  empty - before and after compare equal even though the finding was
-  reported as a change - when the config file's *text* changed while the
-  parsed value did not, e.g. a value that was re-quoted; that case
-  withholds too, saying plainly that the comparison came out empty rather
-  than silently showing nothing, since an empty diff for a reported
-  change would otherwise read as a bug.
+  case-insensitively, since `.claude/settings.json`'s `env` block - the one
+  high-value surface this is the only net for - conventionally holds its keys
+  upper-case. The screen is best-effort with real false negatives: the
+  denylist above is the reliable half.
 
-  A per-report **retention budget** (`RETAIN_BUDGET`, 1 MiB) also bounds
-  the total diff bytes one report may keep. Once it is spent, remaining
-  findings in that report withhold with a note that the budget ran out,
-  and both the report and the exit-time stderr summary add an alarm line:
-  that many changed host-persistence surfaces in one session is treated
-  as a signal in its own right, not merely a limit worth raising - an
-  ordinary session changes few enough of these surfaces that the budget
-  is never in practice a constraint. `diffs = off` in the config's
-  `[audit]` section disables retention (and staging) entirely.
+  A git config key with more than one value on either side is not compared.
+  Git resolves such a key to its last value in FILE order; the audit parses
+  and sorts, so lexicographically-last is not what git runs, and the
+  pre-session side is every value the key had. Comparing anyway produced a
+  diff over values git does not use - two pre-session values read as a
+  deletion that never happened, while the appended value git now executes
+  appeared nowhere. It withholds instead, and must be compared by hand.
+  `include.path` is legitimately multi-valued, so this is an ordinary shape.
 
-  Two residual limits worth knowing on top of the credential screen's
-  false negatives above. The prompt shows at most 40 lines of a diff or
-  of new content before truncating, naming the file so you can read the
-  rest - a large planted change is truncated exactly where you would most
-  want it whole. And the `SIGKILL` case above is a real exposure window,
-  not just a cleanup mechanic: until the next run against that workspace
-  reaps it, the staging directory sits on disk holding whole pre-session
-  file copies (not diffs) of every monitored surface, for as long as the
-  workspace goes unused.
-- Report findings name a path or config key that a session may itself have
-  chosen, so those names reach your terminal as untrusted text. One that
-  contains an ASCII control character, a C1 control, or a Unicode bidi
-  override - which between them can scroll, blank, or visually reorder what
-  you read, and can split a name across the audit's own records - is
-  written hex-encoded behind a `{hex}` marker instead, and decodes with
-  `printf '%b' "$(sed 's/../\\x&/g' <<< HEX)"`. Ordinary names are printed
-  as-is, so the marker appearing at all is itself worth a second look.
-  The encoded form is what `dejima-audit ack` records in the baseline, so
-  what you acknowledge is the file that is actually on disk rather than a
+  A diff can also come out empty when the config file's TEXT changed but the
+  parsed value did not, e.g. a re-quoted value. That withholds too, saying so
+  rather than showing nothing - an empty diff for a reported change reads as
+  a bug.
+
+  A per-report **retention budget** (`RETAIN_BUDGET`, 1 MiB) bounds the total
+  diff bytes one report keeps. Once spent, the rest withhold and both the
+  report and the stderr summary add an alarm line: that many changed
+  host-persistence surfaces in one session is a signal in itself, not a limit
+  worth raising. `diffs = off` in `[audit]` disables retention and staging.
+
+  Two residual limits on top of the screen's false negatives. The prompt
+  truncates at 40 lines, naming the file so you can read the rest - a large
+  planted change is cut exactly where you would want it whole. And the
+  `SIGKILL` case is an exposure window, not just a cleanup mechanic: until
+  the next run reaps it, the staging directory holds whole pre-session file
+  copies of every monitored surface, for as long as the workspace goes
+  unused.
+- Findings name a path or config key the session may have chosen, so they
+  reach your terminal as untrusted text. One containing an ASCII or C1
+  control or a Unicode bidi override - which can scroll, blank or reorder
+  what you read, and split a name across the audit's own records - is written
+  hex-encoded behind a `{hex}` marker and decodes with
+  `printf '%b' "$(sed 's/../\\x&/g' <<< HEX)"`. Ordinary names print as-is,
+  so the marker appearing is itself worth a look. The encoded form is what
+  `ack` records, so what you acknowledge is the file on disk rather than a
   readable approximation of its name.
 - `~/.claude/projects` (session transcripts and per-project memory) is
-  **not** mounted read-only, unlike the code-carrying `~/.claude`
-  directories - deliberately. Scoping a read-only mount to "the
-  current project's slug"
-  would break the sandbox's own worktree workflow: a session running in
-  `WORKSPACE/.worktrees/<branch>` writes to a *different* projects slug
-  than the parent repo, so a slug-scoped overlay would either block that
-  legitimate write or have to be widened until it stopped protecting
-  anything. Instead, cross-project writes are **detected, not prevented**:
-  the exit-time audit above flags any write during the session to another
-  project's `~/.claude/projects/<slug>` directory, excluding the current
-  project's own slug and its worktree children. That's after-the-fact
-  reporting of transcript/memory poisoning in *other* projects, not a
-  block on it happening. Residual even within that: the current project's
-  own transcripts/memory stay fully writable throughout, since Claude Code
-  needs that for host-side `claude --resume` to work; and detection keys
-  off each slug's max file mtime, so an edit to a file that isn't already
-  that project's newest doesn't advance the max and goes unflagged. A
-  read-only mount for `projects/` remains a possible future option if
-  Claude Code is confirmed to fail cleanly (not hang, as `~/.claude.json`
-  did) when a write to it is blocked. This is also the sharpest case of
-  the time-window caveat above: `~/.claude/projects` is a single
-  directory shared by *every* Claude Code session on the host, not scoped
-  to the one being audited, so an entirely innocent concurrent session
-  working on some unrelated project can produce a `CROSS-PROJECT` finding
-  here that reads like the audited session poisoned another project's
-  history - it didn't, it just happened to be running at the same time.
+  **not** mounted as a tree. A session gets the workspace's own slug
+  directory plus one for each live git worktree nested under it,
+  read-write, and nothing else. Every other project's slug is *absent*
+  from the container rather than present-but-unwritable. Agent memory is
+  loaded into context at session start, so a write into another project's
+  `memory/` is text a future session there reads as established background
+  rather than as input - and until #104 the mount alone was enough to do
+  that, with no escape and no credential involved.
+
+  The set is derived from real worktree paths
+  (`bin/dejima-slug-lib.sh`), not from the slug string, because a slug
+  cannot tell a child from a sibling: `/w/proj/wt1` and `/w/proj-wt1`
+  produce the same one. A `<slug>-*` glob therefore also matches any
+  project checked out beside this one whose directory name merely extends
+  it, which is an ordinary layout rather than a contrived one.
+
+  `projects/` itself stays container-local, so a slug that is not mounted
+  is still writable during the session and just does not survive it. That
+  is what lets Claude Code create a slug it does not find, which is the
+  first-run case it already handles, and it is why an absent sibling does
+  not risk the hang that a blocked write to `~/.claude.json` caused.
+
+  Residual: the workspace's own transcripts and memory stay fully writable
+  throughout, since Claude Code needs that for host-side `claude --resume`.
+
+  There is no longer a `CROSS-PROJECT` finding for this. It existed when the
+  whole tree was mounted and a session really could write another project's
+  slug. Now that it cannot, the only writes the audit could still see there
+  are other sessions doing their own work - and on a host running several at
+  once it reported all of them, on every report. Detection that cannot
+  observe what it was built for, and fires constantly on what it was not,
+  trains an operator to skim. The mount is the control, and `static.bats`
+  pins it.
+- `~/.claude/file-history` (Claude Code's edit-undo record) is **not**
+  mounted as a tree either, and the reason is sharper than for `projects/`.
+  Each `<session>/<hash>@v1` is a **verbatim copy of an edited file**, so the
+  tree is other projects' source rather than a record about it. Read access
+  alone is the leak; it was read-write on top of that, and tampering with
+  another project's undo record puts chosen content into a later restore.
+
+  It is keyed by session id, so unlike `projects/` there is no slug in its
+  own paths to scope it by. The transcript names supply it: `projects/<slug>/
+  <id>.jsonl` and `file-history/<id>` name the same session, so the slugs the
+  workspace owns select the session ids it owns
+  (`workspace_session_ids` in `bin/dejima-slug-lib.sh`). A directory whose
+  session has no transcript belongs to no project dejima can identify, and is
+  left out.
+
+  The session about to start has an id nothing can know in advance, so its
+  own directory is created container-local and does not survive the run.
+  That costs undo history for that session if it is later resumed on the
+  host, and it is the reason this cannot be scoped as completely as
+  `projects/`, where the workspace's own slug is known at launch.
+- Epitaphs - the summary a session leaves when its context is compacted away -
+  are kept host-side in `[epitaph] dir`, and that directory is never mounted. A session can neither read what earlier
+  sessions concluded nor edit what this one is about to record, which is the
+  same arrangement as the audit store and for the same reason. The transcript
+  they are extracted from *is* session-writable, so a summary records what was
+  there rather than vouching for it - and nothing loads one back into a later
+  session, deliberately: that would be the channel #104 closed, rebuilt.
 - Allowing egress to a domain at all creates a channel for exfiltrating data
   through *that service's own request surface* (query strings, lookups,
   etc.), not just to a literal attacker-controlled destination. Domain-level
@@ -408,51 +443,45 @@ hands to whichever session starts against it. See
   similar. It shares the host kernel, so a container-escape vulnerability
   would defeat every other control described here at once.
 - Most of what this document asserts is re-checked by the `test/` suite on
-  every change - over 300 bats tests across six files, and `runtime.bats`
-  builds the image and asserts against a real running container rather than
-  a mock: IPv6 blocked, a non-allowlisted domain refused by the DNS proxy
-  and a non-allowlisted destination unreachable, `node` unable to re-run
-  `init-firewall.sh` to widen the allowlist, no working `sudo`. Coverage is
-  not the same as completeness, though. The reasoning in this document is
-  judgement rather than assertions, nothing tests it, and a property nobody
-  thought to write a test for can still regress quietly. A missing `ipset`
-  dedup flag broke firewall startup entirely - twice, on two different
-  ipsets - before anyone caught it; there is a test asserting every
-  `ipset add` is duplicate-safe now, which is the pattern to expect. Each
-  of these tests exists because something was got wrong first.
+  every change, and `runtime.bats` asserts against a real running container
+  rather than a mock: IPv6 blocked, a non-allowlisted domain refused and a
+  non-allowlisted destination unreachable, `node` unable to re-run
+  `init-firewall.sh`, no working `sudo`. Coverage is not completeness: the
+  reasoning here is judgement, nothing tests it, and a property nobody wrote
+  a test for can regress quietly. A missing `ipset` dedup flag broke firewall
+  startup twice before anyone caught it; there is a test for that now, which
+  is the pattern to expect - each of these exists because something was got
+  wrong first.
 - The Anthropic OAuth credential (`~/.claude/.credentials.json`) never
-  enters the container. It simply isn't mounted: `~/.claude` is not shared
-  as a whole (see the mount inventory below), so there is no path from
+  enters the container. It isn't mounted: `~/.claude` is not shared as a
+  whole (see the mount inventory below), so there is no path from
   inside the container to that file. `dejima` starts a host-side proxy
   (`bin/dejima-anthropic-proxy`) on a per-session unix socket and mounts a
-  fake credential in place of the real one - the two secrets are
-  placeholders, while `scopes`/`subscriptionType`/`rateLimitTier`/`clientId`
-  are copied from the real file so the session's own UI stays accurate. Its
-  expiry is set a year out rather than mirrored, deliberately: a near-term
-  one would make Claude Code refresh on its own, with a refresh token that
-  isn't real, and log the session out. Inside the container,
+  fake credential in place of the real one - the secrets are placeholders,
+  while `scopes`/`subscriptionType`/`rateLimitTier`/`clientId` are copied so
+  the session's UI stays accurate. Its expiry is a year out rather than
+  mirrored: a near-term one would make Claude Code refresh on its own with a
+  refresh token that isn't real, logging the session out. Inside the
+  container,
   `entrypoint.sh` points `api.anthropic.com` at a loopback TLS terminator
   (`anthropic-proxy-forwarder.js`) whose certificate is generated at startup
   and trusted only there, and which pipes raw bytes into the proxy's socket.
   The proxy injects the real access token per request and owns OAuth refresh
   - rotated tokens are written back to the real host file atomically, so
   host-side sessions stay logged in. Claude Code itself is configured
-  exactly as it would be on the host: no `ANTHROPIC_BASE_URL`, no
-  `ANTHROPIC_AUTH_TOKEN`, nothing that marks the session as unusual to the
-  client. The proxy then forwards the client's header block verbatim -
-  same names, same casing, same order, repeats intact, and no
-  `Accept-Encoding` it didn't send - changing only the `Authorization`
-  value, and `anthropic-beta` if the OAuth value isn't already there. Left
-  to itself, Python's `http.client` would have rewritten all of that into
-  its own house style and moved the fingerprint from the client to the
-  wire. What it cannot disguise is that the connection is made by CPython's
-  `ssl` rather than by Node: the TLS handshake (JA3/JA4) and the absence of
-  connection reuse still distinguish it at that layer. What remains is
-  session-scoped: a
-  compromised session can still *use* the model API as you while it runs
-  (that's inherent to the sandbox working at all), but it can no longer read
-  a durable credential and exfiltrate it for offline use via the other
-  allowed egress destinations (npm, GitHub, your Gitea host). Caveats:
+  exactly as on the host: no `ANTHROPIC_BASE_URL`, no
+  `ANTHROPIC_AUTH_TOKEN`, nothing marking the session as unusual. The proxy
+  forwards the client's header block verbatim - same names, casing, order and
+  repeats, and no `Accept-Encoding` it didn't send - changing only
+  `Authorization`, and `anthropic-beta` if the OAuth value isn't there. Left
+  alone, Python's `http.client` would rewrite all of that into its own style
+  and move the fingerprint from the client to the wire. It cannot disguise
+  that the connection is CPython's `ssl` rather than Node: the TLS handshake
+  (JA3/JA4) and the absence of connection reuse still distinguish it.
+  What remains is session-scoped: a compromised session can still *use* the
+  API as you while it runs, which is inherent to the sandbox working at all,
+  but it cannot read a durable credential and exfiltrate it for offline use.
+  Caveats:
   `/login` inside the container doesn't work while the proxy is on; the
   refresh flow mirrors a Claude Code internal and may break if upstream
   changes it, in which case the proxy fails loudly (launch abort, 502s on
@@ -506,13 +535,21 @@ hands to whichever session starts against it. See
   are enumerated in `docker-compose.yml`, in three classes, and anything not
   listed is container-local: writable during the session and gone with the
   container.
-  - **Read-write, because it has to survive the session**: `projects/`,
-    `file-history/`, `backups/`, `tasks/`, `plans/`, `sessions/`. All
-    directories, which is what makes them safe (see the detachment note
-    below).
+  - **Read-write, because it has to survive the session**: `backups/`,
+    `tasks/`, `plans/`, `sessions/`, plus the entries of `projects/` and
+    `file-history/` that belong to this workspace. All directories, which is
+    what makes them safe (see the detachment note below). The two trees are
+    not mounted whole: see the `projects/` note further down, and the
+    `file-history/` note below it.
   - **Read-only**: `plugins/`, `hooks/`, `skills/`, `commands/`, `agents/`,
     `output-styles/`, plus the symlink-target directories `dejima`
-    discovers at launch. Hooks execute automatically with no separate
+    discovers at launch. Those targets are discovered rather than
+    configured, so they are checked against the approval store and the
+    audit stores and refused if they overlap either - a hook symlinked into
+    the approval store used to mount it, read-only and silently, and
+    read-only is no protection when the content is every project's
+    approvals. The same check now covers `[mounts] readonly`, which had
+    none at all. Hooks execute automatically with no separate
     approval step, skills can ship executable scripts, and
     commands/agents/output-styles are prompt text loaded into a later host
     session, so without this a compromised session could plant a backdoored
@@ -531,8 +568,7 @@ hands to whichever session starts against it. See
     `statusline.sh` are mounted read-only at `/etc/dejima-claude/` and
     installed into `~/.claude` by `entrypoint.sh` as root-owned, unwritable
     files. A session can still unlink and replace them - `~/.claude` belongs
-    to it - but only its own copy, which was the whole point of making them
-    read-only.
+    to it - but only its own copy, which is why they are read-only.
 
   The reason for the split is that a **single-file bind mount silently
   detaches**. It is pinned to the dentry at its mount point, so when the
@@ -698,6 +734,18 @@ hands to whichever session starts against it. See
   changes takes effect the *next* time `dejima` itself runs,
   not the current one.
 
+  The report splits on whether git can see the change. tcb_snapshot only
+  populates when the workspace IS dejima's own repo, so an ordinary change
+  there is already in the working tree you are about to read in `git status`
+  and in the pull request - reporting each one as a CRITICAL to acknowledge
+  duplicated a stronger control and made developing dejima itself unusable.
+  Those are counted in one INFO line instead.
+
+  A change git *cannot* see stays a CRITICAL finding, because nothing else
+  would catch it: a file in `.gitignore` or `.git/info/exclude`, or one
+  marked `assume-unchanged` or `skip-worktree`, is live code that runs on the
+  next launch and appears nowhere in `git status`.
+
 ### Declaration files are session-writable
 
 A project's `.dejima/` declaration files sit inside the project mount, which
@@ -930,10 +978,9 @@ sessions you have ever run, which is how a Loki instance stops being usable.
 Filter on it with `| json | session=` as above rather than expecting a
 selector.
 
-Two things are absent by design and worth knowing before you go looking for
-them. **Tool results are never shipped**, only invocations, so a query will
-show you that a session ran `Bash` with certain arguments and never what came
-back. **Thinking blocks are never shipped either.** Both dominate the volume
+Two things are absent by design. **Tool results are never shipped**, only
+invocations, so a query will show you that a session ran `Bash` with certain
+arguments and never what came back. **Thinking blocks are never shipped either.** Both dominate the volume
 and neither was what opting in asked for. Long values are clipped, so a very
 large prompt or tool input ends in `...[truncated]` rather than being dropped.
 
@@ -962,6 +1009,32 @@ want GitHub push and PRs as a standing choice instead, set
 `ssh.github.com:22`, but persistently rather than per launch, and uses its
 own `[forge.github] push_token_file` rather than a loaded SSH key.
 
+**Why the default is off.** The principle is narrower than CI and holds
+whether or not CI exists anywhere: *a session that is not being used to push
+to a forge should not be able to push to it.* Your configured forge is the
+one you chose to push to. GitHub is where everything else lives - every
+dependency a session reads, every repository it clones to understand an
+error - so that is where the rule bites by default.
+
+The case is an eager agent, not a compromised one: it finds a real bug in an
+unrelated open-source project and pushes a fix branch to somebody else's
+repository. That this is not an attack does not make it harmless. Maintainers
+are already triaging a flood of unsolicited AI-generated pull requests, and
+one more arriving in your name, unreviewed, is a cost imposed on someone who
+did not ask for it. The block exists so that it cannot happen, not so that
+you get to approve it first.
+
+[Don't point it at repositories you don't own](responsible-use.md#dont-point-it-at-repositories-you-dont-own)
+is the rule this enforces, and gives the fork-and-review workflow to use
+instead: work against your own fork, review the result, and open the upstream
+pull request by hand if you still want to.
+
+Closing the CI chain follows from this - a push is code execution on whatever
+runners that repository can reach, per the section above - but it is not the
+reason for it. Reading it the other way round gets the order backwards, and
+makes the default look like a CI workaround to be turned off once CI is
+handled.
+
 The sandbox can always read from GitHub (clone, fetch, `gh` API calls). By
 **default** it can never push there: enforced at the network layer, not
 just by `git-push-guard.sh`. GitHub's IP ranges are only allowlisted on port

forges/gitea/bin/gitea-actions

@@ -1,23 +1,19 @@
 #!/bin/bash
 # Read Gitea Actions runs and job logs from a shell - the CI-side counterpart
-# to gitea-pr and gitea-issue, and deliberately their twin rather than their
-# sibling: all three are copied into the sandbox image as single
-# self-contained files, and gitea-pr also lives in a dotfiles checkout, so a
-# shared library would be a third thing to keep in step across three places.
-# The duplicated config/token/curl preamble is pinned in all three by
-# test/static.bats instead.
+# to gitea-pr and gitea-issue. Self-contained like both of them rather than
+# sharing a library: all three are copied into the image as single files, so
+# a library would be another thing to keep in step. test/static.bats pins the
+# duplicated config/token/curl preamble instead.
 #
-# Its reason to exist is a sandboxed agent's: CI is where the tests that
-# cannot run in the sandbox actually run, and without this the only way to
-# see why one failed is for the operator to paste the log in by hand.
+# It exists for the agent: CI is where the tests the sandbox cannot run
+# actually run, and without this the only way to see a failure is for the
+# operator to paste the log in.
 #
-# Nothing here touches /repos/{owner}/{repo}/actions/runs, or the job and
-# log endpoints reached through it, however natural they look. Gitea gates
-# that group behind reqOwner(): a token belonging to a dedicated non-owner
-# account - which is what a sandbox should be given - is refused with "user
-# should be the owner of the repo", and no scope grants ownership, so
-# widening the token cannot fix it. The three sources used instead need only
-# read access to the repo:
+# Nothing here touches /repos/{owner}/{repo}/actions/runs or the endpoints
+# reached through it, however natural they look. Gitea gates that group
+# behind reqOwner(), so a dedicated non-owner account - what a sandbox should
+# have - is refused, and no scope grants ownership. The three sources used
+# instead need only read access:
 #
 #   actions/tasks            one record per job, carrying its run number,
 #                            head_sha, event and result

harnesses/claude-code/harness.conf

@@ -1,4 +1,8 @@
-command        = claude
-keys           = projects_dir:path=~/.claude/projects statusline_path:path=~/.claude/statusline.sh
-resume_command = claude --resume {{session_id}}
-briefing       = briefing.md
+command         = claude
+command_path    = /usr/local/share/npm-global/bin/claude
+guards          = bash-deny push-guard merge-guard notify
+no_confirm_args = --dangerously-skip-permissions
+settings_args   = --settings /etc/dejima-settings.json
+keys            = projects_dir:path=~/.claude/projects file_history_dir:path=~/.claude/file-history statusline_path:path=~/.claude/statusline.sh
+resume_command  = claude --resume {{session_id}}
+briefing        = briefing.md

hooks/merge-guard.sh

@@ -1,30 +1,23 @@
 #!/usr/bin/env bash
 # PreToolUse hook: inside dejima, block a merge command matching
-# DEJIMA_MERGE_BLOCK - a literal command prefix (the forge's PR command
-# followed by its merge subcommand) resolved host-side by bin/dejima from
-# the configured forge and workflow
-# (see dejima-provider-lib.sh), not hardcoded to any one forge. The sandbox
-# workflow is feature branch -> open a PR/MR -> a human merges; what lands
-# on the default branch is an operator decision, exactly as it already is
-# for `git push` (see git-push-guard.sh, whose no-direct-push-to-a-blocked-
-# branch rule this is the API-side counterpart to - a merge reaches the
-# branch without a push, so that guard never sees it).
+# DEJIMA_MERGE_BLOCK - a literal command prefix resolved host-side by
+# bin/dejima from the configured forge and workflow, never hardcoded to one
+# forge. What lands on the default branch is an operator decision, as it
+# already is for `git push`; this is the API-side counterpart to
+# git-push-guard.sh, since a merge reaches the branch without a push.
 #
-# Deliberately not conditional on intent. A hook is handed a command, not a
-# conversation, so it cannot tell an instructed merge from an unprompted one;
-# the only honest place to draw the line is the context, and the context that
-# needs the line is this one - headless unattended runs under
-# --dangerously-skip-permissions, where nobody is watching the turn happen.
-# On the host the hook is a no-op and an operator can still ask for a merge.
+# Not conditional on intent: a hook is handed a command, not a conversation,
+# so it cannot tell an instructed merge from an unprompted one. The line is
+# drawn by context instead, and the context that needs it is this one -
+# unattended runs under --dangerously-skip-permissions. On the host it is a
+# no-op.
 #
-# Like its sibling this is a fast-fail guard against an overeager agent, not
-# a hard security boundary: a sandbox session has full shell access and the
-# forge token in its environment, so it could call the merge endpoint with
-# curl directly. It stops the mistake, not the attack.
+# A fast-fail guard against an overeager agent, not a boundary: a session has
+# shell access and the forge token, so it could call the merge endpoint with
+# curl. It stops the mistake, not the attack.
 #
-# DEJIMA_MERGE_BLOCK fails OPEN: empty means no command is blocked, so the
-# hook is inert rather than conditionally installed - a workflow with no
-# merge-block policy (or a forge with nothing to block) genuinely has none.
+# Fails OPEN - empty means nothing is blocked, so a workflow with no
+# merge-block policy genuinely has none rather than the hook guessing.
 
 if [ -z "${DEJIMA:-}" ]; then
     exit 0

hooks/sandbox-context.sh

@@ -66,29 +66,24 @@ fi
 requirements=""
 decl_dir="$project_dir/.dejima"
 
-# Everything below reads attacker-controllable text on its way into an
-# agent's context window, so all three costs it can impose are bounded:
+# Everything below reads attacker-controllable text on its way into a context
+# window, so each cost it can impose is bounded:
 #
 #   MAX_DECLARED  entries considered - the same cap read_declared() applies
-#                 host-side, so this reports on the same window
-#                 `dejima-project approve` will act on.
+#                 host-side, so this reports the window `approve` acts on.
 #   MAX_ENTRY_LEN characters per entry.
+#   MAX_LISTED    entries named; the rest are counted only. Without it, 200
+#                 entries of 200 chars is 40 KB of attacker-chosen tokens per
+#                 section, and a large enough $requirements pushed the jq call
+#                 past MAX_ARG_STRLEN and emitted nothing - costing every
+#                 later session its whole briefing.
 #
-# Within that window the report is a strict subset of what `approve` accepts,
-# never a superset: it never names something approve would reject, but it can
-# stay silent about something approve would take - a dotless domain (the
-# regex below requires a label separator, deliberately, so a bare
-# IGNORE-PREVIOUS-INSTRUCTIONS can't be echoed into the context) and a valid
-# domain longer than MAX_ENTRY_LEN are both approvable host-side and both
-# silently skipped here. Fail-closed in the direction that matters: the worst
-# case is a missing line of advice, not an unapproved entry presented as fine.
-#   MAX_LISTED    entries actually named in the report; the rest are only
-#                 counted. Without it 200 entries of 200 chars would be 40 KB
-#                 of attacker-chosen tokens per section, and a big enough
-#                 $requirements used to push the jq call at the end past
-#                 Linux's 128 KiB MAX_ARG_STRLEN and emit nothing at all -
-#                 silently costing every later session in the project its
-#                 entire briefing.
+# The report is a strict subset of what `approve` accepts, never a superset:
+# it can stay silent about something approvable (a dotless domain, which the
+# regex refuses so a bare IGNORE-PREVIOUS-INSTRUCTIONS cannot be echoed into
+# context; or one longer than MAX_ENTRY_LEN) but never names something
+# `approve` would reject. The worst case is a missing line of advice, not an
+# unapproved entry presented as fine.
 MAX_DECLARED=200
 MAX_ENTRY_LEN=200
 MAX_LISTED=20

install.sh

@@ -1,6 +1,6 @@
 #!/bin/bash
-# Idempotent installer for dejima. Symlinks this repo's hooks into
-# ~/.claude/hooks, registers them in ~/.claude/settings.json, and puts
+# Idempotent installer for dejima. Registers the selected harness's guard
+# hooks wherever that harness reads them from, and puts
 # bin/dejima, bin/dejima-allow, bin/dejima-dep, bin/dejima-project,
 # bin/dejima-audit and the opt-in remote-access scripts on PATH. Safe to
 # re-run - every step is a no-op if already done, and every change is
@@ -35,113 +35,6 @@ fi
 echo "Found required tools: docker, docker compose v2, jq, python3"
 echo
 
-mkdir -p "$HOME/.claude/hooks"
-
-link_hook() {
-    local name="$1" src="$SCRIPT_DIR/hooks/$1" dest="$HOME/.claude/hooks/$1"
-    if [ -L "$dest" ] && [ "$(readlink -f "$dest")" = "$(readlink -f "$src")" ]; then
-        echo "Skipping: $name already linked to this repo's copy"
-        return
-    fi
-    if [ -e "$dest" ] || [ -L "$dest" ]; then
-        mkdir -p "$BACKUP_DIR"
-        echo "Moving: existing ~/.claude/hooks/$name to $BACKUP_DIR/$name"
-        mv "$dest" "$BACKUP_DIR/$name"
-    fi
-    echo "Linking: $name ($src -> $dest)"
-    ln -s "$src" "$dest"
-}
-
-link_hook sandbox-context.sh
-link_hook git-push-guard.sh
-link_hook merge-guard.sh
-
-# Symlink from before the gitea-merge-guard.sh -> merge-guard.sh rename
-# points at a hooks/ path that no longer exists. Only remove it if it's
-# actually dangling - the same check the claude-sandbox-* cleanup below
-# uses for bin/, so an operator's own file at this name is left alone.
-old_merge_guard="$HOME/.claude/hooks/gitea-merge-guard.sh"
-if [ -L "$old_merge_guard" ] && [ ! -e "$old_merge_guard" ]; then
-    echo "Removing: dangling pre-rename symlink $old_merge_guard"
-    rm -f "$old_merge_guard"
-fi
-
-# Linked so it exists at a stable path for the read-only ~/.claude/hooks
-# mount, but deliberately NOT passed to register_hook below: bin/dejima
-# registers it per session, into a throwaway settings copy, and only when
-# remote access is configured. Registering it here would fire it for every
-# unsandboxed host session too, where it has no marker directory to write to.
-link_hook dejima-notify.sh
-
-# deny-dangerous-bash.sh is a general (non-sandbox-gated) hook that may
-# already be owned by another install (e.g. dotfiles) - only claim it if
-# nothing else already provides it, so this repo works fully standalone
-# without clobbering an existing setup.
-deny_dest="$HOME/.claude/hooks/deny-dangerous-bash.sh"
-if [ -e "$deny_dest" ] || [ -L "$deny_dest" ]; then
-    echo "Skipping: deny-dangerous-bash.sh already present at $deny_dest (leaving it as-is)"
-else
-    echo "Linking: deny-dangerous-bash.sh ($SCRIPT_DIR/hooks/deny-dangerous-bash.sh -> $deny_dest)"
-    ln -s "$SCRIPT_DIR/hooks/deny-dangerous-bash.sh" "$deny_dest"
-fi
-echo
-
-SETTINGS="$HOME/.claude/settings.json"
-if [ ! -f "$SETTINGS" ]; then
-    echo "Creating $SETTINGS"
-    echo '{}' > "$SETTINGS"
-fi
-
-# The registration that went with the symlink removed above. Two things
-# depend on it going: Claude Code reports a missing hook command on every
-# session start while it points at a path that no longer exists, and - until
-# the dedupe below stopped matching it - it also stood in for the
-# registration it was renamed from, so `register_hook merge-guard.sh`
-# skipped and the merge guard was registered nowhere at all. Dropped only
-# when the path it names is gone, so an operator's own file at that name
-# keeps its registration, matching the symlink check above.
-if [ ! -e "$old_merge_guard" ] && jq -e --arg s "/gitea-merge-guard.sh" \
-    '[.hooks.PreToolUse[]?.hooks[]?.command // empty] | any(endswith($s))' \
-    "$SETTINGS" >/dev/null 2>&1; then
-    echo "Removing: stale pre-rename gitea-merge-guard.sh registration from $SETTINGS"
-    settings_tmp=$(mktemp)
-    jq --arg s "/gitea-merge-guard.sh" \
-        '.hooks.PreToolUse = ((.hooks.PreToolUse // [])
-            | map(.hooks = ((.hooks // []) | map(select((.command // "") | endswith($s) | not))))
-            | map(select((.hooks | length) > 0)))' \
-        "$SETTINGS" > "$settings_tmp" && mv "$settings_tmp" "$SETTINGS"
-fi
-
-register_hook() {
-    local event="$1" matcher="$2" script_name="$3" description="$4"
-    # endswith on the leading slash, not test(): test() is an unanchored
-    # regex, so a hook whose name merely ends in another's - which is what a
-    # rename like gitea-merge-guard.sh -> merge-guard.sh produces - read as
-    # already registered, and the new hook was silently never added.
-    if jq -e --arg s "/$script_name" \
-        "[.hooks.${event}[]?.hooks[]?.command // empty] | any(endswith(\$s))" \
-        "$SETTINGS" >/dev/null 2>&1; then
-        echo "Skipping: $script_name already registered as a $event hook"
-        return
-    fi
-    echo "Registering $script_name as a $event hook in $SETTINGS ($description)"
-    local tmp
-    tmp=$(mktemp)
-    jq --arg matcher "$matcher" --arg cmd "\$HOME/.claude/hooks/$script_name" \
-        ".hooks.${event} = ((.hooks.${event} // []) + [{\"matcher\": \$matcher, \"hooks\": [{\"type\": \"command\", \"command\": \$cmd}]}])" \
-        "$SETTINGS" > "$tmp" && mv "$tmp" "$SETTINGS"
-}
-
-register_hook SessionStart "startup|resume|clear|compact" sandbox-context.sh \
-    "injects sandbox context on session start"
-register_hook PreToolUse Bash git-push-guard.sh \
-    "blocks pushes outside the allowed host(s)/branches"
-register_hook PreToolUse Bash merge-guard.sh \
-    "blocks the configured forge's merge command inside the sandbox"
-register_hook PreToolUse Bash deny-dangerous-bash.sh \
-    "blocks obviously destructive Bash commands"
-echo
-
 BIN_DIR="$HOME/.local/bin"
 if [[ ":$PATH:" != *":$BIN_DIR:"* ]]; then
     echo "Warning: $BIN_DIR is not on PATH - add it to your shell profile to run 'dejima' by name instead of the full path." >&2
@@ -182,6 +75,7 @@ link_bin dejima-project
 # the startup warning about unacknowledged CRITICAL findings tells the
 # operator to run `dejima-audit ack`, which is only true once it is on PATH.
 link_bin dejima-audit
+link_bin dejima-epitaph
 
 # Remote access (opt-in, see docs/remote-access.md). All three are inert
 # without [remote-access] enabled = on - they refuse to run at all - so
@@ -235,9 +129,9 @@ else
         printf '# dejima configuration. See README.md.\n'
         printf '# Every key below is shown at its default; uncomment to change one.\n'
         printf '#\n'
-        printf '# [core] forge and [core] workflow are the exception: they are written\n'
-        printf '# out uncommented, so this file states what it selects rather than\n'
-        printf '# inheriting whatever a future release defaults to.\n'
+        printf '# [core] harness, forge and workflow are written out uncommented, so\n'
+        printf '# this file states what it selects rather than inheriting a future\n'
+        printf '# default. harness has none at all - dejima refuses to start unset.\n'
         printf '#\n'
         printf '# Those two decide whether this sandbox pushes at all. The shipped\n'
         printf '# default is forge = none with workflow = local-only: the agent\n'
@@ -261,6 +155,7 @@ else
             case "$_key" in
                 forge.gitea.ssh_host)      _seed="${GITEA_SSH_HOST:-}" ;;
                 forge.gitea.api)           _seed="${GITEA_API:-}" ;;
+                core.harness)              _seed="claude-code" ;;
                 core.forge)                _seed="$_starter_forge" ;;
                 core.workflow)             _seed="$_starter_workflow" ;;
                 core.command)              _seed="${SANDBOX_CMD:-}" ;;
@@ -293,6 +188,32 @@ else
     chmod 600 "$CONFIG_FILE"
 fi
 
+# Guard registration, after the starter config above rather than before it:
+# where dejima's guards go, and in what shape, is the selected harness's
+# business, and on a fresh install that selection is the one just seeded.
+# Registering before reading it would mean writing one harness's config for
+# an operator who has chosen another.
+#
+# Delegated to the harness's plugin, which is also what gates each guard on
+# the harness having declared it (see `guards` in its manifest).
+echo
+source "$SCRIPT_DIR/bin/dejima-harness-lib.sh"
+_harness="$(config_get core harness 2>/dev/null)" || _harness=""
+if [ -z "$_harness" ]; then
+    echo "No [core] harness is set in $CONFIG_FILE, so no guard hooks were"
+    echo "registered. Set one and re-run ./install.sh."
+else
+    DEJIMA_HARNESS_NAME="$_harness"
+    harness_plugin_load "$_harness"
+    if declare -F harness_register_guards >/dev/null 2>&1; then
+        echo "Registering $_harness's guard hooks"
+        harness_register_guards "$SCRIPT_DIR" "$BACKUP_DIR"
+    else
+        echo "Harness '$_harness' installs no guard hooks."
+        echo
+    fi
+fi
+
 if [ -f "$(config_root)/remote.conf" ]; then
     echo
     echo "Note: $(config_root)/remote.conf is no longer read - its keys moved to the"

test/audit.bats

@@ -23,6 +23,8 @@ setup() {
     CONFIG_DIR="$(mktemp -d)"
     export DEJIMA_CONFIG_DIR="$CONFIG_DIR"
     cat > "$CONFIG_DIR/config" <<EOF
+[core]
+harness = claude-code
 [harness.claude-code]
 projects_dir = $PROJDIR
 statusline_path = $DEJIMA_AUDIT_STATUSLINE
@@ -143,16 +145,24 @@ cmp_() { "$AUDIT" compare "$WS" "$STATE" 2>&1; }
     [ "$count" -eq 1 ]
 }
 
-@test "flags a cross-project write under ~/.claude/projects" {
+# The cross-project finding is gone, and these are its replacement: writes
+# under ~/.claude/projects must produce NOTHING now, whichever slug they land
+# in. It reported a poisoning route a session could once take - the whole
+# tree was mounted - and since #104 a session gets its own slugs only, so
+# every remaining write it could see belongs to some other session doing its
+# own work. With several sessions running it reported all of them on every
+# report.
+@test "a write to another project's slug is not reported" {
     mkdir -p "$PROJDIR/-some-other-project"
     echo x > "$PROJDIR/-some-other-project/file1"
     snap
     echo y > "$PROJDIR/-some-other-project/file2"; touch -d '+1 hour' "$PROJDIR/-some-other-project/file2"
     run cmp_
-    [[ "$output" == *"CROSS-PROJECT history/memory write: -some-other-project"* ]]
+    [[ "$output" != *"CROSS-PROJECT"* ]]
+    [[ "$output" != *"-some-other-project"* ]]
 }
 
-@test "does not flag the current project's own projects slug" {
+@test "a write to this project's own slug is not reported either" {
     mkdir -p "$PROJDIR/$CURRENT_SLUG"
     echo x > "$PROJDIR/$CURRENT_SLUG/file1"
     snap
@@ -161,15 +171,6 @@ cmp_() { "$AUDIT" compare "$WS" "$STATE" 2>&1; }
     [[ "$output" != *"CROSS-PROJECT"* ]]
 }
 
-@test "does not flag a worktree child's projects slug" {
-    mkdir -p "$PROJDIR/${CURRENT_SLUG}--worktrees-featurex"
-    echo x > "$PROJDIR/${CURRENT_SLUG}--worktrees-featurex/file1"
-    snap
-    echo y > "$PROJDIR/${CURRENT_SLUG}--worktrees-featurex/file2"; touch -d '+1 hour' "$PROJDIR/${CURRENT_SLUG}--worktrees-featurex/file2"
-    run cmp_
-    [[ "$output" != *"CROSS-PROJECT"* ]]
-}
-
 @test "silent when no projects changed" {
     mkdir -p "$PROJDIR/-some-other-project"
     echo x > "$PROJDIR/-some-other-project/file1"
@@ -211,7 +212,11 @@ cmp_() { "$AUDIT" compare "$WS" "$STATE" 2>&1; }
     [[ "$output" == *".claude/evil"* ]]
 }
 
-@test "TCB: flags a change under bin/ when workspace is the audit's own repo root" {
+@test "TCB: a change git can see is reported as info, not as a CRITICAL to ack" {
+    # tcb_snapshot only populates when the workspace IS dejima's own repo, so
+    # every change here is also in the git working tree the operator is about
+    # to read. One CRITICAL per file duplicated `git status` and made
+    # developing dejima itself unusable.
     mkdir -p "$WS/bin"
     echo x > "$WS/bin/x"
     echo x > "$WS/install.sh"
@@ -219,7 +224,27 @@ cmp_() { "$AUDIT" compare "$WS" "$STATE" 2>&1; }
     snap
     echo changed > "$WS/bin/x"
     run cmp_
-    [[ "$output" == *"TCB (sandbox self) modified: bin/x"* ]]
+    [[ "$output" != *"TCB (sandbox self) modified"* ]]   # nothing to acknowledge
+    # Recorded as INFO in the report rather than shouted on stderr: the
+    # startup warning already said the TCB is writable this session.
+    slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
+    run bash -c "cat '$AUDITDIR/$slug'/*.txt"
+    [[ "$output" == *"this workspace IS dejima"* ]]
+}
+
+@test "TCB: a change git CANNOT see is still a CRITICAL finding" {
+    # The case that justifies keeping the check at all. A file listed in
+    # .git/info/exclude is live code that runs on the next launch and appears
+    # nowhere in `git status`, so nothing but this hash sees it.
+    mkdir -p "$WS/bin" "$WS/.git/info"
+    echo x > "$WS/bin/hidden"
+    echo x > "$WS/install.sh"
+    echo 'bin/hidden' > "$WS/.git/info/exclude"
+    export DEJIMA_AUDIT_REPO_ROOT="$(readlink -f "$WS")"
+    snap
+    echo changed > "$WS/bin/hidden"
+    run cmp_
+    [[ "$output" == *"TCB (sandbox self) modified, INVISIBLE to git: bin/hidden"* ]]
 }
 
 @test "TCB: does not flag changes when the workspace is not the audit's own repo root" {
@@ -241,6 +266,24 @@ cmp_() { "$AUDIT" compare "$WS" "$STATE" 2>&1; }
     [ "$output" = "" ]
 }
 
+@test "says so when no harness is selected, rather than auditing a narrower surface in silence" {
+    # The statusline path comes from the selected harness, so with none
+    # selected there is nothing to watch and the compensating control for
+    # the symlink gap covers nothing. bin/dejima refuses to start without
+    # [core] harness, so this state means the audit was driven by hand - and
+    # a clean report over a surface that was never examined is worse than a
+    # noisy one.
+    printf '#!/bin/sh\necho original\n' > "$DEJIMA_AUDIT_STATUSLINE"
+    cat > "$CONFIG_DIR/config" <<EOF
+[audit]
+dir = $AUDITDIR
+baselines_dir = $BASEDIR
+EOF
+    run "$AUDIT" snapshot "$WS" "$STATE"
+    [[ "$output" == *"no [core] harness is set"* ]] || { echo "$output"; return 1; }
+    [[ "$output" == *"NOT checked"* ]] || { echo "$output"; return 1; }
+}
+
 @test "flags a plain regular-file statusline.sh replaced with different content" {
     printf '#!/bin/sh\necho original\n' > "$DEJIMA_AUDIT_STATUSLINE"
     snap
@@ -592,14 +635,56 @@ cmp_() { "$AUDIT" compare "$WS" "$STATE" 2>&1; }
     [ "$result" -ne 124 ]
 }
 
-@test "show prints the only report there is, and names it" {
+# One `ack` run used to walk one report and exit, saying nothing about the
+# rest, so clearing a backlog took as many runs as there were reports and
+# looked like the tool was stuck on the first one.
+@test "ack walks every report in one run when told to continue" {
+    slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
+    snap; echo 'export X=1' > "$WS/.envrc"; cmp_ >/dev/null 2>&1
+    snap; mkdir -p "$WS/.claude"; echo '{}' > "$WS/.claude/settings.json"; cmp_ >/dev/null 2>&1
+    [ "$(find "$AUDITDIR/$slug" -name '*.txt' | wc -l)" -eq 2 ]
+    # approve, continue, approve
+    run bash -c "printf 'a\ny\na\n' | '$AUDIT' ack '$WS'"
+    [ "$status" -eq 0 ]
+    [[ "$output" == *"2 reports waiting"* ]]
+    [[ "$output" == *".envrc"* ]]
+    [[ "$output" == *"settings.json"* ]]      # the second report, same run
+}
+
+@test "ack stops at one report when told not to continue" {
+    slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
+    snap; echo 'export X=1' > "$WS/.envrc"; cmp_ >/dev/null 2>&1
+    snap; mkdir -p "$WS/.claude"; echo '{}' > "$WS/.claude/settings.json"; cmp_ >/dev/null 2>&1
+    run bash -c "printf 'a\nn\n' | '$AUDIT' ack '$WS'"
+    [ "$status" -eq 0 ]
+    [[ "$output" == *"still waiting"* ]]      # said what is left behind
+    [[ "$output" != *"settings.json"* ]]      # and did not open it
+}
+
+@test "show indexes the reports rather than dumping one" {
+    # It used to `cat` a single report: a wall of text that answered neither
+    # question an operator has - how much is waiting, and what still needs a
+    # decision - and included findings already acknowledged, so it did not
+    # even shrink as they worked through it.
     snap
     echo 'export X=1' > "$WS/.envrc"
     cmp_ >/dev/null 2>&1
     run bash -c "'$AUDIT' show '$WS' 2>&1"
     [ "$status" -eq 0 ]
+    [[ "$output" == *"1 report(s)"* ]]
+    [[ "$output" == *"-CRITICAL.txt"* ]]      # names it
+    [[ "$output" == *"ack starts here"* ]]    # and which one ack will open
+    [[ "$output" != *".envrc"* ]]             # without the contents
+}
+
+@test "show --full prints the report ack would open" {
+    snap
+    echo 'export X=1' > "$WS/.envrc"
+    cmp_ >/dev/null 2>&1
+    run bash -c "'$AUDIT' show '$WS' --full 2>&1"
+    [ "$status" -eq 0 ]
     [[ "$output" == *".envrc"* ]]
-    [[ "$output" == *"-CRITICAL.txt"* ]]   # said which report it printed
+    [[ "$output" == *"-CRITICAL.txt"* ]]
 }
 
 @test "show and ack agree on which report they mean" {
@@ -619,19 +704,20 @@ WARNING  changed workspace file: .vscode/settings.json
 REPORT
     run "$AUDIT" show "$WS"
     [ "$status" -eq 0 ]
-    [[ "$output" == *".envrc"* ]]           # the stale CRITICAL report...
+    [[ "$output" == *"-CRITICAL.txt"* ]]    # the stale CRITICAL report...
     [[ "$output" != *"settings.json"* ]]    # ...not the newer clean one
 }
 
-@test "show falls back to the newest report when no CRITICAL one is pending" {
+@test "show marks the newest report as the one ack will open, absent a CRITICAL" {
     slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
     report_dir="$AUDITDIR/$slug"; mkdir -p "$report_dir"
     echo 'older' > "$report_dir/2020-01-01-000000.txt"
     echo 'newer' > "$report_dir/2026-01-01-000001.txt"
     run "$AUDIT" show "$WS"
     [ "$status" -eq 0 ]
-    [[ "$output" == *"newer"* ]]
-    [[ "$output" != *"older"* ]]
+    [[ "$output" == *"2 report(s)"* ]]             # both listed, not just one
+    [[ "$output" == *"2020-01-01-000000.txt"* ]]
+    [[ "$output" == *"2026-01-01-000001.txt"*"ack starts here"* ]]
 }
 
 @test "ack presents a stale unacknowledged CRITICAL report before a newer clean one" {
@@ -772,39 +858,32 @@ REPORT
     [[ "$output" != *"-CRITICAL.txt"* ]]
 }
 
-@test "a finding that cannot be acknowledged is presented once, not on every run" {
-    # There is nothing to record for one, so without a marker it re-prompted
-    # forever and `ack` could never reach a state of having nothing to ask.
+@test "a cleared finding is presented once, not on every run" {
+    # There is nothing to record for a surface that no longer exists, so
+    # without a `reviewed` marker the report re-prompted forever and `ack`
+    # could never reach a state of having nothing to ask.
     slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
-    mkdir -p "$PROJDIR/-some-other-project"
-    echo x > "$PROJDIR/-some-other-project/file1"
-    snap
-    echo y > "$PROJDIR/-some-other-project/file2"
-    touch -d '+1 hour' "$PROJDIR/-some-other-project/file2"
-    cmp_ >/dev/null 2>&1
-    run bash -c "printf 's\n' | '$AUDIT' ack '$WS'"
-    [[ "$output" == *"CROSS-PROJECT"* ]]
+    snap; echo 'export X=1' > "$WS/.envrc"; cmp_ >/dev/null 2>&1
+    rm -f "$WS/.envrc"
+    run bash -c "printf 'a\n' | '$AUDIT' ack '$WS'"
+    [[ "$output" == *".envrc"* ]]
     grep -q '^reviewed ' "$BASEDIR/$slug"
 
-    run bash -c "printf 's\n' | '$AUDIT' ack '$WS'"
-    [[ "$output" != *"CROSS-PROJECT"* ]]
+    run bash -c "printf 'a\n' | '$AUDIT' ack '$WS'"
+    [[ "$output" != *".envrc"* ]]
 }
 
 @test "quitting a report does not mark it reviewed" {
-    # The marker means "walked to the end". An abort discards everything
-    # else, and must discard this too or one q would silence the report.
+    # The marker means "walked to the end". An abort discards everything else,
+    # and must discard this too or one q would silence the report.
     slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
-    mkdir -p "$PROJDIR/-some-other-project"
-    echo x > "$PROJDIR/-some-other-project/file1"
-    snap
-    echo y > "$PROJDIR/-some-other-project/file2"
-    touch -d '+1 hour' "$PROJDIR/-some-other-project/file2"
-    cmp_ >/dev/null 2>&1
+    snap; echo 'export X=1' > "$WS/.envrc"; cmp_ >/dev/null 2>&1
+    rm -f "$WS/.envrc"
     run bash -c "printf 'q\n' | '$AUDIT' ack '$WS'"
     ! grep -q '^reviewed ' "$BASEDIR/$slug" 2>/dev/null
 
-    run bash -c "printf 's\n' | '$AUDIT' ack '$WS'"
-    [[ "$output" == *"CROSS-PROJECT"* ]]
+    run bash -c "printf 'a\n' | '$AUDIT' ack '$WS'"
+    [[ "$output" == *".envrc"* ]]
 }
 
 @test "a WARNING git config finding can be acknowledged, both ways (#45)" {
@@ -854,24 +933,6 @@ REPORT
     [[ "$output" != *"-CRITICAL.txt"* ]]
 }
 
-@test "ack does not write an unmatchable entry for a finding with no id" {
-    # Was written against a changed git config, which #45 made ackable. The
-    # invariant is about findings that have no id to record at all, so it now
-    # uses the one class that still has none: a cross-project write, whose
-    # subject is another project's slug rather than anything this audit can
-    # hash and pin.
-    slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
-    mkdir -p "$PROJDIR/-some-other-project"
-    echo x > "$PROJDIR/-some-other-project/file1"
-    snap
-    echo y > "$PROJDIR/-some-other-project/file2"
-    touch -d '+1 hour' "$PROJDIR/-some-other-project/file2"
-    cmp_ >/dev/null 2>&1
-    run bash -c "printf 'a\n' | '$AUDIT' ack '$WS'"
-    [[ "$output" == *"cannot"* ]]
-    ! grep -qE '^(safe|spam) ' "$BASEDIR/$slug" 2>/dev/null
-}
-
 @test "an ack-approved finding is suppressed on the next compare (id round-trips through the baseline)" {
     snap; echo 'export X=1' > "$WS/.envrc"; cmp_ >/dev/null 2>&1
     run bash -c "printf 'a\n' | '$AUDIT' ack '$WS'"
@@ -900,24 +961,37 @@ REPORT
                                              # so the report must stay eviction-protected
 }
 
-@test "ack refuses to approve a finding whose path no longer exists, and the CRITICAL suffix survives" {
-    # A since-deleted path has no content left to pin a `safe` approval to -
-    # writing one anyway (the old behaviour) produced a baseline entry
-    # baseline_suppresses() can never match (its own empty-sha guard rejects
-    # it as malformed), so the surface stayed permanently unsuppressible -
-    # while `a` still advanced crit_acked and let the report shed its
-    # -CRITICAL.txt protection for good. Refusing outright keeps both the
-    # baseline and the report's protection honest.
+@test "a finding whose path no longer exists is cleared, not refused forever" {
+    # Refusing outright kept the baseline honest - there is no content left to
+    # pin a `safe` entry to - but it also meant crit_acked never reached
+    # crit_seen, so the report kept -CRITICAL.txt. select_report prefers
+    # CRITICAL, so one deleted file pinned `ack` to a report that no decision
+    # could resolve and hid every other report behind it, permanently.
+    #
+    # Still nothing written to the baseline: an entry with an empty sha can
+    # never match. The finding is simply cleared, because a surface that does
+    # not exist does not run. If it comes back, it comes back as new.
     slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
     snap; echo 'export X=1' > "$WS/.envrc"; cmp_ >/dev/null 2>&1
     rm -f "$WS/.envrc"   # gone by the time ack runs
     run bash -c "printf 'a\n' | '$AUDIT' ack '$WS'"
     [ "$status" -eq 0 ]
-    [[ "$output" == *"cannot approve: .envrc no longer exists"* ]]
-    [[ "$output" == *"nothing recorded"* ]]
-    [ ! -f "$BASEDIR/$slug" ]                    # nothing written to the baseline at all
+    [[ "$output" == *"cleared"* ]]
+    [[ "$output" == *".envrc"* ]]
+    ! grep -qE '^(safe|spam) ' "$BASEDIR/$slug" 2>/dev/null   # nothing unmatchable recorded
+    grep -q '^reviewed ' "$BASEDIR/$slug"                     # but the report is done with
     run bash -c "ls '$AUDITDIR/$slug'"
-    [[ "$output" == *"-CRITICAL.txt"* ]]         # unacknowledged CRITICAL keeps its protection
+    [[ "$output" != *"-CRITICAL.txt"* ]]                      # and no longer blocks the queue
+}
+
+# The other half: clearing is not a way to launder a finding whose surface is
+# still there. Only an absent one can be cleared.
+@test "a finding whose path still exists is approved against its content, not cleared" {
+    slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
+    snap; echo 'export X=1' > "$WS/.envrc"; cmp_ >/dev/null 2>&1
+    run bash -c "printf 'a\n' | '$AUDIT' ack '$WS'"
+    [[ "$output" != *"cleared"* ]]
+    grep -q '^safe .envrc ' "$BASEDIR/$slug"
 }
 
 @test "a workspace path containing its own ': ' cannot hijack ack's id derivation" {
@@ -1060,7 +1134,10 @@ REPORT
 
 @test "ack records a safe tcb entry whose id round-trips and suppresses the next compare" {
     slug="$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
-    mkdir -p "$WS/bin"; echo x > "$WS/bin/x"; echo x > "$WS/install.sh"
+    # Hidden from git, which is the only shape that is still a finding.
+    mkdir -p "$WS/bin" "$WS/.git/info"
+    echo x > "$WS/bin/x"; echo x > "$WS/install.sh"
+    echo 'bin/x' > "$WS/.git/info/exclude"
     export DEJIMA_AUDIT_REPO_ROOT="$(readlink -f "$WS")"
     snap
     echo changed > "$WS/bin/x"
@@ -1118,7 +1195,7 @@ REPORT
     cmp_ >/dev/null
     run bash -c "cd '$WS' && '$AUDIT' show 2>/dev/null"
     [ "$status" -eq 0 ]
-    [[ "$output" == *".envrc"* ]]
+    [[ "$output" == *"report(s) for $WS"* ]]
 }
 
 @test "show finds the workspace when run from a subdirectory" {
@@ -1128,7 +1205,7 @@ REPORT
     cmp_ >/dev/null
     run bash -c "cd '$WS/src/deep' && '$AUDIT' show 2>/dev/null"
     [ "$status" -eq 0 ]
-    [[ "$output" == *".envrc"* ]]
+    [[ "$output" == *"report(s) for $WS"* ]]
 }
 
 @test "show finds the parent repo when run from a linked worktree" {
@@ -1143,7 +1220,7 @@ REPORT
     cmp_ >/dev/null
     run bash -c "cd '$WS/.worktrees/wt' && '$AUDIT' show 2>/dev/null"
     [ "$status" -eq 0 ]
-    [[ "$output" == *".envrc"* ]]
+    [[ "$output" == *"report(s) for $WS"* ]]
 }
 
 @test "show finds the parent repo's report even after the worktree itself was snapshotted" {
@@ -1158,7 +1235,7 @@ REPORT
     "$AUDIT" snapshot "$WS/.worktrees/wt" "$(mktemp)"
     run bash -c "cd '$WS/.worktrees/wt' && '$AUDIT' show 2>/dev/null"
     [ "$status" -eq 0 ]
-    [[ "$output" == *".envrc"* ]]
+    [[ "$output" == *"report(s) for $WS"* ]]
 }
 
 @test "show names the path it resolved when that workspace has no reports" {
@@ -1179,7 +1256,7 @@ REPORT
     other="$(mktemp -d)"; git -C "$other" init -q
     run bash -c "cd '$other' && '$AUDIT' show '$WS' 2>/dev/null"
     [ "$status" -eq 0 ]
-    [[ "$output" == *".envrc"* ]]
+    [[ "$output" == *"report(s) for $WS"* ]]
     rm -rf "$other"
 }
 
@@ -1324,9 +1401,9 @@ key_of() { printf '%s' "$1" | sha256sum | cut -d' ' -f1; }
     snap
     head -1 "$STATE" | grep -q '^### staging /'
     # The existing section ranges are anchored at ### gitconfig and
-    # ### statusline, so putting staging first must leave them untouched.
+    # ### hostexec, so putting staging first must leave them untouched.
     grep -q '^### gitconfig$' "$STATE"
-    grep -q '^### statusline$' "$STATE"
+    grep -q '^### hostexec$' "$STATE"
 }
 
 @test "denylisted surfaces are never staged, not even transiently" {
@@ -1853,3 +1930,43 @@ sidecar() { local r; r="$(find "$AUDITDIR" -name '*.txt' | sort | tail -1)"; r="
     run "$AUDIT" compare "$WS" "$STATE"
     [[ "$output" != *"writable"* ]]
 }
+
+@test "a host-execution finding still records the id existing baselines hold" {
+    # The id is what an approval is stored under. Generalising the mechanism
+    # must not rename it: every baseline an operator already has says
+    # "statusline:statusline.sh", and an id that no longer matches re-fires
+    # a CRITICAL they approved. The namespace and leaf come from the plugin
+    # for exactly this reason - they are the harness's to keep stable.
+    printf '#!/bin/sh\necho original\n' > "$DEJIMA_AUDIT_STATUSLINE"
+    snap
+    printf '#!/bin/sh\necho changed\n' > "$DEJIMA_AUDIT_STATUSLINE"
+    run cmp_
+    [[ "$output" == *"HOST-CODE-EXECUTION"* ]] || { echo "$output"; return 1; }
+    # Approve it, then confirm the baseline line is the historical id.
+    printf 'a\n' | "$AUDIT" ack "$WS" >/dev/null 2>&1 || true
+    run cat "$BASEDIR/$(printf '%s' "$WS" | sed 's/[^a-zA-Z0-9]/-/g')"
+    [[ "$output" == *"safe statusline:statusline.sh "* ]] \
+        || { echo "baseline id changed: $output"; return 1; }
+}
+
+@test "a harness declaring no host-execution surface is watched for none" {
+    # Not every harness runs something on the host from its own config. The
+    # absence must mean nothing is checked, not that the previous harness's
+    # path is.
+    printf '#!/bin/sh\necho original\n' > "$DEJIMA_AUDIT_STATUSLINE"
+    snap
+    cat > "$CONFIG_DIR/config" <<EOF
+[core]
+harness = claude-code
+[harness.claude-code]
+projects_dir = $PROJDIR
+statusline_path =
+[audit]
+dir = $AUDITDIR
+baselines_dir = $BASEDIR
+EOF
+    printf '#!/bin/sh\necho changed\n' > "$DEJIMA_AUDIT_STATUSLINE"
+    run cmp_
+    [[ "$output" != *"HOST-CODE-EXECUTION"* ]] \
+        || { echo "watched a surface the harness declares no path for: $output"; return 1; }
+}

test/config.bats

@@ -188,6 +188,16 @@ EOF
 
 write_config() {
     printf '%s\n' "$@" > "$CONFIG_DIR/config"
+    # A launch is refused without a harness, so every fixture needs one. Its
+    # own [core] section, which is legal and resolves the same; a test that
+    # means to omit it writes the file directly.
+    # A key assignment, not the word anywhere in the arguments: a [mounts]
+    # path containing "harness" would otherwise suppress the seeding.
+    local _a _has=0
+    for _a in "$@"; do
+        [[ "$_a" =~ ^[[:space:]]*harness[[:space:]]*= ]] && _has=1
+    done
+    [ "$_has" -eq 1 ] || printf '[core]\nharness = claude-code\n' >> "$CONFIG_DIR/config"
 }
 
 @test "config_get returns the schema default when the key is absent" {
@@ -1045,24 +1055,16 @@ run_install() {
 
     local name out hit=0
 
-    # docs/ and systemd/ are pure operator-facing prose - nothing there is
-    # live shell plumbing, so any mention of a retired name is either the
-    # deliberate "no longer read" notice (install.sh's remote.conf one lives
-    # outside these two directories, but the exclusion is kept here too in
-    # case a future doc adds one), a stale instruction, or one of the two
-    # cases below that are neither:
+    # docs/ and systemd/ are operator-facing prose, no live plumbing, so a
+    # retired name there is either the deliberate "no longer read" notice, a
+    # stale instruction, or one of two cases that are neither:
     #
-    # - A manifest-field line (`env_export    = GITEA_API=api ...`): this is
-    #   docs/providers.md quoting a forge manifest's `env_export` value
-    #   verbatim, which *declares* a container-side environment variable -
-    #   the exact thing env_export is for - not an instruction to an operator
-    #   to set one. `GITEA_API` and friends are retired only as host-side,
-    #   operator-set variables (what _warn_legacy_env's table is about); they
-    #   are not retired as the container-side names env_export produces, and
-    #   docs/providers.md documenting that mapping is not a regression.
-    # - A backtick-quoted mention (`` `GITEA_API` ``): prose naming a
-    #   variable - "the container gets `GITEA_API`" - reads a value, it
-    #   doesn't tell anyone to set one.
+    # - A manifest-field line (`env_export = GITEA_API=api ...`), which is
+    #   providers.md quoting a manifest verbatim. These names are retired as
+    #   host-side operator-set variables, not as the container-side names
+    #   env_export produces.
+    # - A backtick-quoted mention - prose naming a variable reads a value, it
+    #   does not tell anyone to set one.
     while IFS= read -r name; do
         out="$(grep -rnE "\b${name}\b" "$REPO_ROOT/docs" "$REPO_ROOT/systemd" 2>/dev/null \
             | grep -v 'no longer read' \
@@ -1134,11 +1136,273 @@ exit 0
 EOF
     chmod +x "$STUB_DIR/docker"
     docker_log="$(mktemp)"
-    HOME="$RW_HOME" PATH="$STUB_DIR:$PATH" SSH_AUTH_SOCK= \
+    HOME="$RW_HOME" PATH="$STUB_DIR:$PATH" SSH_AUTH_SOCK="${RW_AGENT_SOCK-}" \
         SANDBOX_TEST_DOCKER_LOG="$docker_log" \
         run "$REPO_ROOT/bin/dejima" "$1"
 }
 
+# A real agent, carrying nothing. `ssh-add -l` against it exits nonzero,
+# which is the condition the "no identities loaded" warning keys on, and a
+# handmade socket file would not reproduce it.
+_empty_agent() {
+    eval "$(ssh-agent -s)" >/dev/null
+    RW_AGENT_SOCK="$SSH_AUTH_SOCK"
+    RW_AGENT_PID="$SSH_AGENT_PID"
+}
+_kill_agent() { [ -n "${RW_AGENT_PID:-}" ] && kill "$RW_AGENT_PID" 2>/dev/null; return 0; }
+
+# The forwarded agent is the fallback; [core] ssh_key is the arrangement the
+# docs recommend, because it is a dedicated identity and works unattended.
+# So an empty agent is only worth mentioning when nothing else can push -
+# otherwise the warning fires at exactly the operators who took the advice.
+# ~/.claude/file-history is keyed by session id, so nothing in its own paths
+# says which project a directory belongs to. The transcript names are what
+# resolve it, and this is the assertion that the resolution is right: a
+# session of this workspace is mounted, one belonging to another project is
+# not, and neither is one whose slug merely extends this workspace's - the
+# sibling case that #104 turned on.
+@test "file-history is mounted for this workspace's sessions and no others" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    local slug sib own foreign sibling
+    slug="$(printf '%s' "$workspace" | sed 's/[^a-zA-Z0-9]/-/g')"
+    sib="$slug-extra"
+    own=11111111-1111-1111-1111-111111111111
+    foreign=22222222-2222-2222-2222-222222222222
+    sibling=33333333-3333-3333-3333-333333333333
+    mkdir -p "$RW_HOME/.claude/projects/$slug" \
+             "$RW_HOME/.claude/projects/-some-other-project" \
+             "$RW_HOME/.claude/projects/$sib"
+    : > "$RW_HOME/.claude/projects/$slug/$own.jsonl"
+    : > "$RW_HOME/.claude/projects/-some-other-project/$foreign.jsonl"
+    : > "$RW_HOME/.claude/projects/$sib/$sibling.jsonl"
+    mkdir -p "$RW_HOME/.claude/file-history/$own" \
+             "$RW_HOME/.claude/file-history/$foreign" \
+             "$RW_HOME/.claude/file-history/$sibling"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch "$workspace"
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    grep -qF "file-history/$own" "$docker_log" \
+        || { echo "own session not mounted:"; grep file-history "$docker_log"; return 1; }
+    run grep -F "file-history/$foreign" "$docker_log"
+    [ "$status" -ne 0 ] || { echo "another project's file-history was mounted"; return 1; }
+    run grep -F "file-history/$sibling" "$docker_log"
+    [ "$status" -ne 0 ] || { echo "a sibling project's file-history was mounted"; return 1; }
+    rm -rf "$workspace"
+}
+
+# No default on purpose: a release changing it would move an operator who
+# never chose to a different harness on upgrade.
+@test "a config with no [core] harness is refused, and says what to add" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    # Written directly, not through write_config, which seeds a harness for
+    # every other fixture precisely because a launch needs one.
+    printf '[core]\nssh_key = %s\nforge = none\nworkflow = local-only\n' \
+        "$RW_KEY" > "$CONFIG_DIR/config"
+    _rw_launch "$workspace"
+    [ "$status" -ne 0 ] || { echo "launched with no harness selected:"; echo "$output"; return 1; }
+    [[ "$output" == *"[core] harness is not set"* ]]
+    [[ "$output" == *"harness = claude-code"* ]]   # the line to add
+    [[ "$output" == *"Available:"*"claude-code"* ]]
+    rm -rf "$workspace"
+}
+
+@test "the harness schema carries no default to fall back to" {
+    # Directly again: write_config seeds one, which is the whole point of it.
+    printf '[core]\nforge = none\nworkflow = local-only\n' > "$CONFIG_DIR/config"
+    run lib config_get core harness
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    [ "$output" = "" ]
+}
+
+# --print-config is what an operator reaches for to find out what is unset, so
+# it has to survive the thing being unset.
+@test "--print-config still works when no harness is selected" {
+    printf '[core]\nforge = none\nworkflow = local-only\n' > "$CONFIG_DIR/config"
+    run "$REPO_ROOT/bin/dejima" --print-config
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    [[ "$output" == *"core.harness="* ]]
+    [[ "$output" != *"invalid harness name"* ]]
+}
+
+@test "an empty ssh-agent is not warned about when [core] ssh_key is set" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _empty_agent
+    _rw_launch "$workspace"
+    _kill_agent
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    [[ "$output" != *"no identities loaded"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+# The positive control for the test above: with no key configured the
+# forwarded agent really is the only way to push, and an empty one is worth
+# saying so before it fails deep into a session.
+# --extra-readonly / --extra-read-write append to the config's own lists
+# rather than carrying their own mount code, so these assert the mount lands
+# and then that every refusal the configured form gets applies to the flag
+# too. If they ever stop sharing a path, the guard tests below pass while the
+# flags quietly stop being covered by them.
+_rw_launch_flags() {
+    local ws="$1"; shift
+    STUB_DIR="$(mktemp -d)"
+    cat > "$STUB_DIR/docker" <<'EOF'
+#!/bin/sh
+for a in "$@"; do printf '%s\n' "$a" >> "$SANDBOX_TEST_DOCKER_LOG"; done
+exit 0
+EOF
+    chmod +x "$STUB_DIR/docker"
+    docker_log="$(mktemp)"
+    HOME="$RW_HOME" PATH="$STUB_DIR:$PATH" SSH_AUTH_SOCK= \
+        SANDBOX_TEST_DOCKER_LOG="$docker_log" \
+        run "$REPO_ROOT/bin/dejima" "$@" "$ws"
+}
+
+@test "--extra-readonly and --extra-read-write mount for one launch" {
+    _rw_setup
+    workspace="$(mktemp -d)"; ro="$(mktemp -d)"; rw="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch_flags "$workspace" --extra-readonly "$ro" --extra-read-write "$rw"
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    grep -qxF "$ro:$ro:ro" "$docker_log" || { cat "$docker_log"; return 1; }
+    grep -qxF "$rw:$rw" "$docker_log" || { cat "$docker_log"; return 1; }
+    rm -rf "$workspace" "$ro" "$rw"
+}
+
+@test "the short forms are the same flags, and both are repeatable" {
+    _rw_setup
+    workspace="$(mktemp -d)"; a="$(mktemp -d)"; b="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch_flags "$workspace" --ro "$a" --ro "$b"
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    grep -qxF "$a:$a:ro" "$docker_log" || { cat "$docker_log"; return 1; }
+    grep -qxF "$b:$b:ro" "$docker_log" || { cat "$docker_log"; return 1; }
+    rm -rf "$workspace" "$a" "$b"
+}
+
+# The flags add to the configured lists; they do not replace them. A launch
+# that dropped the config's own entries would still pass the two tests above.
+@test "a flag adds to [mounts] readonly rather than replacing it" {
+    _rw_setup
+    workspace="$(mktemp -d)"; configured="$(mktemp -d)"; extra="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only' \
+                 '[mounts]' "readonly = $configured"
+    _rw_launch_flags "$workspace" --ro "$extra"
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    grep -qxF "$configured:$configured:ro" "$docker_log" || { cat "$docker_log"; return 1; }
+    grep -qxF "$extra:$extra:ro" "$docker_log" || { cat "$docker_log"; return 1; }
+    rm -rf "$workspace" "$configured" "$extra"
+}
+
+# The point of appending to the same list: a flag is a way to say where, not
+# a way past the refusals. Both directions, since the two lists are guarded
+# by different code.
+@test "--extra-readonly is refused when it names the approval store" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch_flags "$workspace" --ro "$CONFIG_DIR"
+    [ "$status" -ne 0 ] || { echo "launched anyway:"; echo "$output"; return 1; }
+    [[ "$output" == *"Refusing --extra-readonly path"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+@test "--extra-read-write is refused when it names the approval store" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch_flags "$workspace" --rw "$CONFIG_DIR"
+    [ "$status" -ne 0 ] || { echo "launched anyway:"; echo "$output"; return 1; }
+    [[ "$output" == *"Refusing --extra-read-write path"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+# A refusal that named "[mounts] readonly" would send someone to a config
+# file that does not mention the path.
+@test "a rejected flag path is reported as a flag, not as a config entry" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch_flags "$workspace" --ro /nonexistent-flag-path
+    [ "$status" -ne 0 ]
+    [[ "$output" == *"--extra-readonly path does not exist"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+# Both lists are colon-separated, so such a path cannot be represented in one
+# and would split into two mounts that each fail the existence check.
+@test "a flag path containing a colon is refused" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch_flags "$workspace" --ro /tmp/has:colon
+    [ "$status" -ne 0 ]
+    [[ "$output" == *"must not contain a colon"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+# #15. Both of these reach `extra_mounts+=(-v ...)` with a path computed at
+# run time, so the static grep for a literal $PROJECT_STORE / config_root
+# cannot see either one - it passed throughout. [mounts] read_write has been
+# guarded since it was added; these two were not.
+@test "[mounts] readonly is refused when it names the approval store" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only' \
+                 '[mounts]' "readonly = $CONFIG_DIR"
+    _rw_launch "$workspace"
+    [ "$status" -ne 0 ] || { echo "launched anyway:"; echo "$output"; return 1; }
+    [[ "$output" == *"Refusing [mounts] readonly entry"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+# The target is discovered from a symlink rather than configured, so nothing
+# had vetted it: a hook pointing into the approval store mounted the store.
+@test "a ~/.claude symlink is refused when its target sits in the approval store" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    mkdir -p "$RW_HOME/.claude/hooks" "$CONFIG_DIR/sub"
+    : > "$CONFIG_DIR/sub/hook.sh"
+    ln -s "$CONFIG_DIR/sub/hook.sh" "$RW_HOME/.claude/hooks/hook.sh"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch "$workspace"
+    [ "$status" -ne 0 ] || { echo "launched anyway:"; echo "$output"; return 1; }
+    [[ "$output" == *"Refusing ~/.claude symlink target"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+# The ordinary case the symlink scan exists for: this repo's own hooks/ are
+# symlinked into ~/.claude and must still be mounted. Without this, refusing
+# everything would pass the two tests above.
+@test "a ~/.claude symlink to an ordinary directory is still mounted" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    hookdir="$(mktemp -d)"
+    : > "$hookdir/hook.sh"
+    mkdir -p "$RW_HOME/.claude/hooks"
+    ln -s "$hookdir/hook.sh" "$RW_HOME/.claude/hooks/hook.sh"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch "$workspace"
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    grep -qxF "$hookdir:$hookdir:ro" "$docker_log" \
+        || { echo "hook target not mounted:"; cat "$docker_log"; return 1; }
+    rm -rf "$workspace" "$hookdir"
+}
+
+@test "an empty ssh-agent is warned about when [core] ssh_key is not set" {
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' 'forge = none' 'workflow = local-only'
+    _empty_agent
+    _rw_launch "$workspace"
+    _kill_agent
+    [[ "$output" == *"no identities loaded"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
 @test "a read_write entry is mounted without :ro, unlike a readonly one" {
     _rw_setup
     rw="$(mktemp -d)"
@@ -1338,7 +1602,7 @@ EOF
     # a gitea token reaching the container: the shipped defaults select no
     # forge at all, so leaving them implicit would mean nothing ever reads
     # token_file and the refusal under test could never fire.
-    printf '[core]\nfirewall = on\nssh_key = %s\nforge = gitea\nworkflow = worktree-pr\n[forge.gitea]\nssh_host = git.example\napi = https://api.example\ntoken_file = %s/token\n' \
+    printf '[core]\nharness = claude-code\nfirewall = on\nssh_key = %s\nforge = gitea\nworkflow = worktree-pr\n[forge.gitea]\nssh_host = git.example\napi = https://api.example\ntoken_file = %s/token\n' \
         "$RW_KEY" "$CONFIG_DIR" > "$CONFIG_DIR/config"
     printf 'real-secret\ngarbage\n' > "$CONFIG_DIR/token"
     STUB_DIR="$(mktemp -d)"
@@ -1357,3 +1621,250 @@ EOF
         || { echo "docker was invoked despite the refusal"; return 1; }
     rm -rf "$workspace" "$RW_HOME" "$STUB_DIR"
 }
+
+# --- epitaph: what a session leaves when its context dies ---------------------------------------------------
+#
+# What a long session worked out is in the summary Claude Code writes when it
+# compacts, and that goes away with the container. The store is host-side and
+# never mounted, so a later session can neither read what an earlier one
+# concluded nor edit what this one is about to record (#104).
+
+_epitaph_fixture() {
+    EPI_STORE="$(mktemp -d)"
+    EPI_TR="$(mktemp -d)"
+    EPI_WS="$(mktemp -d)"; git -C "$EPI_WS" init -q
+    EPI_SLUG="$(printf '%s' "$EPI_WS" | sed 's/[^a-zA-Z0-9]/-/g')"
+    mkdir -p "$EPI_TR/$EPI_SLUG"
+    write_config '[epitaph]' "dir = $EPI_STORE"
+}
+
+# One compact_boundary carrying the metadata, then the summary itself.
+_epitaph_transcript() {
+    local f="$EPI_TR/$EPI_SLUG/$1.jsonl" text="${2:-worked out X}" uid="${3:-aaaabbbb}"
+    {
+        printf '{"type":"system","subtype":"compact_boundary","compactMetadata":{"trigger":"manual","preTokens":900,"postTokens":90}}\n'
+        printf '{"type":"user","isCompactSummary":true,"uuid":"%s","sessionId":"s-1","timestamp":"2026-01-02T03:04:05.000Z","message":{"content":"%s"}}\n' "$uid" "$text"
+    } > "$f"
+}
+
+@test "an epitaph is written out of the container's reach" {
+    _epitaph_fixture
+    _epitaph_transcript s1 "the thing this session worked out"
+    run "$REPO_ROOT/bin/dejima-epitaph" save "$EPI_WS" "$EPI_TR"
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    local saved; saved="$(find "$EPI_STORE" -name '*.md')"
+    [ -n "$saved" ]
+    grep -q 'the thing this session worked out' "$saved"
+    grep -q 'trigger: manual' "$saved"
+    grep -q '900 before, 90 after' "$saved"
+    rm -rf "$EPI_STORE" "$EPI_TR" "$EPI_WS"
+}
+
+@test "saving twice does not duplicate an epitaph already kept" {
+    # `save` runs on every exit, and a session compacted three times has two
+    # epitaphs an earlier run already wrote.
+    _epitaph_fixture
+    _epitaph_transcript s1 "once only"
+    "$REPO_ROOT/bin/dejima-epitaph" save "$EPI_WS" "$EPI_TR" >/dev/null 2>&1
+    "$REPO_ROOT/bin/dejima-epitaph" save "$EPI_WS" "$EPI_TR" >/dev/null 2>&1
+    [ "$(find "$EPI_STORE" -name '*.md' | wc -l)" -eq 1 ]
+    rm -rf "$EPI_STORE" "$EPI_TR" "$EPI_WS"
+}
+
+@test "another project's epitaphs are not collected" {
+    _epitaph_fixture
+    _epitaph_transcript s1 "mine"
+    mkdir -p "$EPI_TR/-some-other-project"
+    printf '{"type":"user","isCompactSummary":true,"uuid":"cccc","sessionId":"s-2","timestamp":"2026-01-02T03:04:05.000Z","message":{"content":"theirs"}}\n' \
+        > "$EPI_TR/-some-other-project/s2.jsonl"
+    "$REPO_ROOT/bin/dejima-epitaph" save "$EPI_WS" "$EPI_TR" >/dev/null 2>&1
+    run grep -rl 'theirs' "$EPI_STORE"
+    [ "$status" -ne 0 ] || { echo "collected another project's summary"; return 1; }
+    grep -rq 'mine' "$EPI_STORE"
+    rm -rf "$EPI_STORE" "$EPI_TR" "$EPI_WS"
+}
+
+@test "no epitaph is written when [epitaph] enabled is off" {
+    _epitaph_fixture
+    write_config '[epitaph]' "dir = $EPI_STORE" 'enabled = off'
+    _epitaph_transcript s1 "should not be kept"
+    "$REPO_ROOT/bin/dejima-epitaph" save "$EPI_WS" "$EPI_TR" >/dev/null 2>&1
+    [ -z "$(find "$EPI_STORE" -name '*.md' 2>/dev/null)" ]
+    rm -rf "$EPI_STORE" "$EPI_TR" "$EPI_WS"
+}
+
+@test "a slug that kept nothing gets no directory" {
+    # A workspace's worktrees are slugs of its own, and they have transcripts;
+    # what they mostly do not have is a compaction summary in any of them. A
+    # directory per slug therefore made the store read as one entry per
+    # worktree, all of them empty.
+    _epitaph_fixture
+    _epitaph_transcript s1 "kept"
+    git -C "$EPI_WS" -c user.email=t@example.com -c user.name=t \
+        commit -q --allow-empty -m init
+    local wt="$EPI_WS/wt" wt_slug
+    git -C "$EPI_WS" worktree add -q "$wt" -b side
+    wt_slug="$(printf '%s' "$wt" | sed 's/[^a-zA-Z0-9]/-/g')"
+    mkdir -p "$EPI_TR/$wt_slug"
+    printf '{"type":"user","message":{"content":"no summary in here"}}\n' \
+        > "$EPI_TR/$wt_slug/s2.jsonl"
+    # A directory an earlier run already left behind goes too.
+    mkdir -p "$EPI_STORE/$wt_slug"
+    "$REPO_ROOT/bin/dejima-epitaph" save "$EPI_WS" "$EPI_TR" >/dev/null 2>&1
+    [ ! -d "$EPI_STORE/$wt_slug" ] \
+        || { echo "empty directory left for $wt_slug"; return 1; }
+    [ -n "$(find "$EPI_STORE/$EPI_SLUG" -name '*.md' 2>/dev/null)" ] \
+        || { echo "the workspace's own epitaph was not kept"; return 1; }
+    git -C "$EPI_WS" worktree remove --force "$wt" 2>/dev/null || true
+    rm -rf "$EPI_STORE" "$EPI_TR" "$EPI_WS"
+}
+
+@test "the epitaph store is not mounted into the container" {
+    # The whole point: a session must not be able to read what earlier
+    # sessions concluded, or edit what this one is about to record.
+    run grep -n 'epitaph' "$REPO_ROOT/docker/docker-compose.yml"
+    [ "$status" -ne 0 ] || { echo "$output"; return 1; }
+    run grep -nE 'extra_mounts\+=.*epitaph' "$REPO_ROOT/bin/dejima"
+    [ "$status" -ne 0 ] || { echo "$output"; return 1; }
+}
+
+# --- harness plugins ------------------------------------------------------
+
+@test "a harness with a manifest but no plugin is refused" {
+    # Support is per harness and deliberate. A manifest alone would present
+    # as a supported harness while none of what dejima does for a harness
+    # had been written, which is the "thin drop-in manifest" this design
+    # exists to prevent.
+    _rw_setup
+    workspace="$(mktemp -d)"
+    # A provider root of its own, via the DEJIMA_PROVIDER_ROOT test hook,
+    # rather than a directory added to the checkout: a suite that writes into
+    # the repository it is testing leaves litter when it aborts and collides
+    # with a concurrent run.
+    proot="$(mktemp -d)"
+    cp -a "$REPO_ROOT/forges" "$REPO_ROOT/workflows" "$REPO_ROOT/harnesses" "$proot/"
+    mkdir -p "$proot/harnesses/test-no-plugin"
+    printf 'command = true\n' > "$proot/harnesses/test-no-plugin/harness.conf"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only' \
+                 'harness = test-no-plugin'
+    DEJIMA_PROVIDER_ROOT="$proot" _rw_launch "$workspace"
+    [ "$status" -ne 0 ] || { echo "launched anyway:"; echo "$output"; return 1; }
+    [[ "$output" == *"No harness plugin for 'test-no-plugin'"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace" "$proot"
+}
+
+@test "emit_mount refuses a harness state directory that overlaps a guarded store" {
+    # The reason the refusal lives in emit_mount rather than in each plugin.
+    # A harness's state paths are computed from config rather than written
+    # down, so this is exactly the shape that reaches a mount without anyone
+    # having looked at the value - the same blind spot as #15.
+    _rw_setup
+    workspace="$(mktemp -d)"
+    audit_dir="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only' \
+                 '[audit]' "dir = $audit_dir" \
+                 '[harness.claude-code]' "projects_dir = $audit_dir/projects"
+    _rw_launch "$workspace"
+    [ "$status" -ne 0 ] || { echo "launched anyway:"; echo "$output"; return 1; }
+    [[ "$output" == *"Refusing harness state mount"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace" "$audit_dir"
+}
+
+@test "a harness that cannot host a guard says so at launch" {
+    # The capability bar allows a harness to be supported at a smaller guard
+    # set. What it does not allow is that being invisible: an operator who
+    # has these everywhere else would otherwise have to read a manifest to
+    # find out they are gone here.
+    _rw_setup
+    workspace="$(mktemp -d)"
+    proot="$(mktemp -d)"
+    cp -a "$REPO_ROOT/forges" "$REPO_ROOT/workflows" "$REPO_ROOT/harnesses" "$proot/"
+    # Same harness, minus two guards, so the message is the only difference.
+    sed -i 's/^guards .*/guards          = bash-deny notify/' \
+        "$proot/harnesses/claude-code/harness.conf"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    DEJIMA_PROVIDER_ROOT="$proot" _rw_launch "$workspace"
+    [[ "$output" == *"does not support 2 of dejima's guards"* ]] || { echo "$output"; return 1; }
+    [[ "$output" == *"push-guard"* ]] || { echo "$output"; return 1; }
+    [[ "$output" == *"merge the agent"* || "$output" == *"merge its own pull requests"* ]] \
+        || { echo "$output"; return 1; }
+    rm -rf "$workspace" "$proot"
+}
+
+@test "a harness hosting every guard says nothing about guards" {
+    # The message must not become a line an operator learns to scroll past.
+    _rw_setup
+    workspace="$(mktemp -d)"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    _rw_launch "$workspace"
+    [[ "$output" != *"guards"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace"
+}
+
+@test "a harness declaring a guard dejima does not know is refused" {
+    # A typo would otherwise turn a guard off silently, which is the failure
+    # the declaration exists to prevent.
+    _rw_setup
+    workspace="$(mktemp -d)"
+    proot="$(mktemp -d)"
+    cp -a "$REPO_ROOT/forges" "$REPO_ROOT/workflows" "$REPO_ROOT/harnesses" "$proot/"
+    sed -i 's/^guards .*/guards          = bash-deny push-gaurd/' \
+        "$proot/harnesses/claude-code/harness.conf"
+    write_config '[core]' "ssh_key = $RW_KEY" 'forge = none' 'workflow = local-only'
+    DEJIMA_PROVIDER_ROOT="$proot" _rw_launch "$workspace"
+    [ "$status" -ne 0 ] || { echo "launched anyway:"; echo "$output"; return 1; }
+    [[ "$output" == *"unknown guard: push-gaurd"* ]] || { echo "$output"; return 1; }
+    rm -rf "$workspace" "$proot"
+}
+
+@test "install.sh registers the selected harness's guards, gated on its declarations" {
+    # Which guards exist and how they are installed is the harness's, so a
+    # harness declaring fewer gets fewer - and dejima does not write a
+    # config belonging to a harness the operator did not choose.
+    local proot; proot="$(mktemp -d)"
+    cp -a "$REPO_ROOT/forges" "$REPO_ROOT/workflows" "$REPO_ROOT/harnesses" "$proot/"
+    sed -i 's/^guards .*/guards          = bash-deny/' \
+        "$proot/harnesses/claude-code/harness.conf"
+    run_install DEJIMA_PROVIDER_ROOT="$proot"
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    run jq -r '[.hooks.PreToolUse[]?.hooks[]?.command] | join(" ")' \
+        "$CONFIG_DIR/.claude/settings.json"
+    [[ "$output" == *"deny-dangerous-bash.sh"* ]] || { echo "$output"; return 1; }
+    [[ "$output" != *"git-push-guard.sh"* ]] \
+        || { echo "registered an undeclared guard: $output"; return 1; }
+    [[ "$output" != *"merge-guard.sh"* ]] \
+        || { echo "registered an undeclared guard: $output"; return 1; }
+    rm -rf "$proot"
+}
+
+@test "install.sh registers every guard when the harness declares them all" {
+    # The shipped harness, unmodified: the behaviour this replaced, kept.
+    run_install
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    local cmds
+    cmds="$(jq -r '[.hooks.PreToolUse[]?.hooks[]?.command,
+                    .hooks.SessionStart[]?.hooks[]?.command] | join(" ")' \
+        "$CONFIG_DIR/.claude/settings.json")"
+    local h
+    for h in sandbox-context.sh git-push-guard.sh merge-guard.sh deny-dangerous-bash.sh; do
+        [[ "$cmds" == *"$h"* ]] || { echo "$h not registered: $cmds"; return 1; }
+    done
+    # Linked for the read-only hooks mount, but registered per session by
+    # bin/dejima rather than here, where it would fire for host sessions too.
+    [ -L "$CONFIG_DIR/.claude/hooks/dejima-notify.sh" ]
+    [[ "$cmds" != *"dejima-notify.sh"* ]] \
+        || { echo "notify registered host-wide: $cmds"; return 1; }
+}
+
+@test "install.sh registers no guards when no harness is selected" {
+    # Writing one harness's config for an operator who has chosen another,
+    # or none, is the coupling this removed.
+    # A config that already exists is never overwritten, so this is how an
+    # operator upgrading from before [core] harness was required arrives.
+    printf '[core]\nforge = none\nworkflow = local-only\n' > "$CONFIG_DIR/config"
+    run_install
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    [[ "$output" == *"No [core] harness is set"* ]] || { echo "$output"; return 1; }
+    [ ! -f "$CONFIG_DIR/.claude/settings.json" ] \
+        || { echo "wrote a harness config anyway"; return 1; }
+}

test/project.bats

@@ -92,7 +92,8 @@ launch() {
     local conf
     conf="$(bash -c 'source "$1"; config_path' _ "$LIB")"
     mkdir -p "$(dirname "$conf")"
-    printf '[core]\nssh_key = %s\n' "$LAUNCH_DIR/sandbox-key" >> "$conf"
+    # No default, so a config that launches must select one.
+    printf '[core]\nharness = claude-code\nssh_key = %s\n' "$LAUNCH_DIR/sandbox-key" >> "$conf"
     HOME="$LAUNCH_DIR/home" PATH="$LAUNCH_DIR:$PATH" \
         SANDBOX_TEST_RUN_ENV="$LAUNCH_ENV" \
         SANDBOX_TEST_DOCKER_LOG="$LAUNCH_DIR/docker-log" \
@@ -817,7 +818,7 @@ ENTRYPOINT [\"/usr/local/bin/entrypoint.sh\"]" ]]
     # no domains at all, or by the env prefix below never reaching the child.
     printf 'good.example.com\n' > "$CONFIG_DIR/allowed-domains"
     launch_setup
-    printf '[core]\nssh_key = %s\n' "$LAUNCH_DIR/sandbox-key" >> "$CONFIG_DIR/config"
+    printf '[core]\nharness = claude-code\nssh_key = %s\n' "$LAUNCH_DIR/sandbox-key" >> "$CONFIG_DIR/config"
     HOME="$LAUNCH_DIR/home" PATH="$LAUNCH_DIR:$PATH" \
         SANDBOX_TEST_RUN_ENV="$LAUNCH_ENV" \
         SSH_AUTH_SOCK= ALLOWED_DOMAINS=evil.example \
@@ -908,6 +909,42 @@ ENTRYPOINT [\"/usr/local/bin/entrypoint.sh\"]" ]]
     [ "$(launched_env SANDBOX_IMAGE)" = "dejima:$(id -un)-$slug" ]
 }
 
+# nvim's data/state/cache are Docker volumes, not host paths, so nothing a
+# session installs reaches the host's own nvim. They do outlive the container
+# though, and under one fixed name each they were shared by every project on
+# the machine: a Lua file dropped under lazy/, or a binary replaced under
+# mason/bin/, is executed by whatever project's session next runs nvim in the
+# sandbox. Scoped per project, the same way the derived image tag is.
+@test "the nvim volumes are per project, not one set shared by every project" {
+    launch_setup
+    run launch "$PROJ"
+    [ "$status" -eq 0 ]
+    slug="$(bash -c 'source "$1"; project_slug "$2"' _ "$LIB" "$PROJ")"
+    [ "$(launched_env SANDBOX_NVIM_DATA)" = "dejima_nvim-data$slug" ]
+    [ "$(launched_env SANDBOX_NVIM_STATE)" = "dejima_nvim-state$slug" ]
+    [ "$(launched_env SANDBOX_NVIM_CACHE)" = "dejima_nvim-cache$slug" ]
+
+    local other; other="$(mktemp -d)"
+    run launch "$other"
+    [ "$status" -eq 0 ]
+    [ "$(launched_env SANDBOX_NVIM_DATA)" != "dejima_nvim-data$slug" ]
+    rm -rf "$other"
+}
+
+# project_slug turns a leading / into -, and a Docker volume name may not
+# start with one: without the prefix every launch would fail at `compose run`
+# rather than at anything that names the cause.
+@test "the derived nvim volume names are valid Docker volume names" {
+    launch_setup
+    run launch "$PROJ"
+    [ "$status" -eq 0 ]
+    local v
+    for v in SANDBOX_NVIM_DATA SANDBOX_NVIM_STATE SANDBOX_NVIM_CACHE; do
+        [[ "$(launched_env "$v")" =~ ^[a-zA-Z0-9][a-zA-Z0-9_.-]*$ ]] \
+            || { echo "$v=$(launched_env "$v") is not a valid volume name"; return 1; }
+    done
+}
+
 @test "the launcher builds and runs the global layer when only global deps are approved" {
     "$REPO_ROOT/bin/dejima-dep" add --global apt:libpq-dev >/dev/null
     launch_setup

test/providers.bats

@@ -236,7 +236,8 @@ assert_remedy_forges_exist() {
     local forge workflow harness
     forge="$(lib config_default core.forge)"
     workflow="$(lib config_default core.workflow)"
-    harness="$(lib config_default core.harness)"
+    # No default to read: the shipped pair under test is forge/workflow.
+    harness=claude-code
     [ "$(lib manifest_get workflow "$workflow" push)" = "off" ] \
         || { echo "default workflow '$workflow' pushes - a fresh install has no forge to push to"; return 1; }
     run lib provider_validate_selection "$forge" "$workflow" "$harness"
@@ -591,6 +592,80 @@ assert_remedy_forges_exist() {
     [[ "$output" != *"create"*"PR"* ]] || { echo "branch-push briefing mentions opening a PR"; return 1; }
 }
 
+# #32: an implementer ran a bare `git config user.name ...` inside a
+# worktree and wrote the parent checkout's real config, misattributing three
+# commits. A worktree isolates the working tree, not `.git/config`, and the
+# briefing told the agent to use worktrees while saying nothing about that -
+# which reads as more isolation than is on offer. The same unscoped command
+# reaches core.hooksPath and credential.helper, which execute on the host.
+#
+# In the core paragraph rather than a workflow's own prose, so a workflow can
+# opt into worktrees but cannot ship them without this.
+@test "every worktree briefing says a worktree does not isolate .git/config" {
+    local wf
+    for wf in worktree-pr branch-push; do
+        [ "$(lib manifest_get workflow "$wf" requires_worktree)" = "on" ]
+        run lib briefing_compose gitea "$wf" claude-code
+        [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+        [[ "$output" == *".git/config"* ]] \
+            || { echo "$wf briefing does not mention .git/config"; return 1; }
+        [[ "$output" == *"git -C"* ]] \
+            || { echo "$wf briefing does not give the scoped form"; return 1; }
+    done
+}
+
+# The negative half: a workflow that does not use worktrees has no reason to
+# carry the paragraph, and its presence there would mean the splice is
+# unconditional rather than opt-in.
+@test "a workflow without worktrees does not carry the worktree paragraph" {
+    [ "$(lib manifest_get workflow local-only requires_worktree)" = "off" ]
+    run lib briefing_compose none local-only claude-code
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    [[ "$output" != *"does not isolate"* ]]
+    [[ "$output" != *"worktree MUST live"* ]]
+}
+
+# The default that blocks GitHub push exists to stop an eager agent pushing to
+# somebody else's project. When the operator lifts it, the agent gains the
+# ability without the reason, so the briefing supplies the reason.
+@test "the briefing warns against pushing elsewhere when GitHub push is on" {
+    DEJIMA_GITHUB_PUSH=on run lib briefing_compose gitea worktree-pr claude-code
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    [[ "$output" == *"GitHub push is enabled for this session"* ]]
+    [[ "$output" == *"pointed at and nothing else"* ]]
+    [[ "$output" == *"not a contribution"* ]]
+    # and it says what to do instead of nothing at all
+    [[ "$output" == *"say so in your reply"* ]]
+}
+
+# Two points from docs/responsible-use.md that an agent can act on, as
+# opposed to the majority which are the operator's to follow. Kept short on
+# purpose: this text is in every session's context, for every user.
+@test "the briefing names the two responsible-use points an agent can act on" {
+    run lib briefing_compose gitea worktree-pr claude-code
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    # a domain that acts on the world rather than serving data
+    [[ "$output" == *"action rather than a source"* ]]
+    [[ "$output" == *"let the operator decide"* ]]
+    # and things in the working tree that are not code
+    [[ "$output" == *"not"*"code"* ]]
+    [[ "$output" == *".env"* ]]
+}
+
+@test "the briefing says nothing about GitHub push when it is off" {
+    run lib briefing_compose gitea worktree-pr claude-code
+    [ "$status" -eq 0 ] || { echo "$output"; return 1; }
+    [[ "$output" != *"GitHub push is enabled"* ]]
+}
+
+@test "bin/dejima turns the GitHub push briefing on for both routes to it" {
+    # --allow-github-push, and forge = github under a pushing workflow.
+    grep -q 'DEJIMA_GITHUB_PUSH=on' "$REPO_ROOT/bin/dejima"
+    grep -q 'ALLOW_GITHUB_PUSH_FLAG" = 1 \]; then' "$REPO_ROOT/bin/dejima"
+    grep -q 'hosts_allowlisted_by_core" ) \?= "on"\|hosts_allowlisted_by_core)" = "on"' "$REPO_ROOT/bin/dejima"
+    grep -q 'export DEJIMA_GITHUB_PUSH' "$REPO_ROOT/bin/dejima"
+}
+
 @test "local-only pushes nothing and blocks no branches" {
     run lib manifest_validate workflow local-only
     [ "$status" -eq 0 ] || { echo "$output"; return 1; }

test/proxy.bats

@@ -409,21 +409,16 @@ make_cert() {
     # shell *function* persists in the calling shell rather than scoping to
     # the call, so it would leak into every later test in this file.
     #
-    # The -u pair is what makes this test about the code rather than about
-    # where it runs (#56). "No cert configured" was expressed by not setting
-    # the variables, which only holds where nothing else set them -
-    # entrypoint.sh:211 exports both into every tls-transport session, so a
-    # suite run from inside a dejima container inherited real cert paths, the
-    # guard below never fired, and execution fell through to the listen. What
-    # failed then was an EADDRINUSE against the session's own forwarder on
-    # :443, reported as this assertion failing - the code was fine and the
-    # test was reading its own environment. Unsetting them states the
-    # precondition instead of assuming it.
+    # The -u pair makes this about the code rather than where it runs (#56).
+    # "No cert configured" used to be expressed by not setting the variables,
+    # which only holds where nothing else did - entrypoint.sh exports both
+    # into every tls-transport session, so a suite run inside a dejima
+    # container inherited real paths, the guard never fired, and the
+    # fall-through hit EADDRINUSE on :443 and was reported as this assertion
+    # failing. Unsetting states the precondition instead of assuming it.
     #
-    # The port override is belt-and-braces: the cert check runs before the
-    # listen today, so a passing test never binds anything, but if that order
-    # ever changes the fall-through lands on an unused port rather than on
-    # whatever holds :443.
+    # The port override is belt-and-braces: the cert check precedes the listen
+    # today, so a fall-through lands on an unused port if that ever changes.
     run env -u SANDBOX_ANTHROPIC_TLS_CERT -u SANDBOX_ANTHROPIC_TLS_KEY \
         SANDBOX_ANTHROPIC_PROXY_SOCKET="$SOCK" SANDBOX_ANTHROPIC_TRANSPORT=tls \
         SANDBOX_ANTHROPIC_PROXY_PORT=18398 \

test/remote.bats

@@ -5,6 +5,25 @@
 # compromised session writes into - so most of what's here is about what it
 # refuses to do with that input.
 
+# No test in this file may modify the repository's own CLIs. Two of them need
+# an exit-time audit that fails, and they get it from a temp shadow of bin/
+# (see shadow_repo) rather than by overwriting the real one. This is what
+# proves it stayed that way even if a test aborts partway through, and it is
+# checked against a hash rather than against `git status` so that it holds
+# while someone has uncommitted edits to that file.
+setup_file() {
+    REMOTE_AUDIT_SUM="$(sha256sum "$BATS_TEST_DIRNAME/../bin/dejima-audit" | cut -d' ' -f1)"
+    export REMOTE_AUDIT_SUM
+}
+teardown_file() {
+    local now
+    now="$(sha256sum "$BATS_TEST_DIRNAME/../bin/dejima-audit" | cut -d' ' -f1)"
+    [ "$now" = "$REMOTE_AUDIT_SUM" ] || {
+        echo "test/remote.bats modified bin/dejima-audit in the repository" >&2
+        return 1
+    }
+}
+
 setup() {
     REPO_ROOT="$BATS_TEST_DIRNAME/.."
     LIB="$REPO_ROOT/bin/dejima-project-lib.sh"
@@ -19,15 +38,6 @@ teardown() {
     if [ -n "${TMUX_DIR:-}" ]; then
         TMUX_TMPDIR="$TMUX_DIR" tmux kill-server 2>/dev/null || true
     fi
-    # Safety net for tests that temporarily swap out the real bin/dejima-
-    # audit: each such test restores it itself right after `run`, before any
-    # assertion that could short-circuit the test body, but this catches the
-    # case anyway if something in between still managed to abort first.
-    if [ -n "${AUDIT_BACKUP:-}" ] && [ -f "$AUDIT_BACKUP" ]; then
-        cp "$AUDIT_BACKUP" "$REPO_ROOT/bin/dejima-audit"
-        chmod +x "$REPO_ROOT/bin/dejima-audit"
-        rm -f "$AUDIT_BACKUP"
-    fi
     rm -rf "$CONFIG_DIR" "$RUNTIME_DIR" ${STUB_DIR:+"$STUB_DIR"} ${LAUNCH_DIR:+"$LAUNCH_DIR"} \
         ${TMUX_DIR:+"$TMUX_DIR"} ${FAKE_HOME:+"$FAKE_HOME"}
 }
@@ -70,6 +80,54 @@ write_remote_config() {
     [ "$output" = "$RUNTIME_DIR/dejima" ]
 }
 
+# A throwaway copy of the repository whose bin/dejima-audit a test may
+# replace or delete. Everything else is a symlink to the real thing, so this
+# costs one directory of links.
+#
+# bin/ is copied rather than linked, and both reasons matter. bin/dejima
+# resolves its own location with `readlink -f`, so a symlinked copy would
+# point SCRIPT_DIR back at the repository and the stub would never be the one
+# it calls. And a symlink is actively dangerous here: `cat > $SHADOW/bin/
+# dejima-audit` follows it and truncates the real script, which is precisely
+# what this helper exists to stop. Copying removes the whole class rather
+# than one instance of it.
+#
+# Two tests here need an audit binary that fails, and they used to get one by
+# overwriting `$REPO_ROOT/bin/dejima-audit` and restoring it afterwards. That
+# left the working tree holding a two-line stub for the length of the test,
+# which is a poor thing to do to a security tool: a run killed partway, or
+# two runs of this file at once, restores the stub permanently instead of the
+# script, and a suite that passes while silently disabling the exit-time
+# audit is the exact failure this file exists to prevent. It also, reasonably
+# enough, reads like an attack to anything watching the repository for
+# changes.
+shadow_repo() {
+    local root entry
+    root="$(cd "$REPO_ROOT" && pwd)"
+    SHADOW="$BATS_TEST_TMPDIR/shadow"
+    mkdir -p "$SHADOW"
+    for entry in "$root"/*; do
+        [ "$(basename "$entry")" = bin ] && continue
+        ln -sfn "$entry" "$SHADOW/$(basename "$entry")"
+    done
+    cp -a "$root/bin" "$SHADOW/bin"
+    DEJIMA_BIN="$SHADOW/bin/dejima"
+}
+
+# Stands in for the real audit and does nothing but fail. Self-describing
+# because a file watcher can surface it mid-test, and a bare `exit 3` sitting
+# where a security tool belongs is alarming without this comment on it.
+shadow_failing_audit() {
+    cat > "$SHADOW/bin/dejima-audit" <<'EOF'
+#!/bin/bash
+# Test fixture for test/remote.bats, living in that test's temp directory
+# and never in the repository. Fails on purpose: bin/dejima must still
+# report the container's exit code and not this one.
+exit 3
+EOF
+    chmod +x "$SHADOW/bin/dejima-audit"
+}
+
 # A fake `docker` that records its argv, so bin/dejima's launch path can be
 # exercised with no daemon. Same approach as test/project.bats.
 stub_launch() {
@@ -119,12 +177,13 @@ launch() {
     local conf
     conf="$(bash -c 'source "$1"; config_path' _ "$LIB")"
     mkdir -p "$(dirname "$conf")"
-    printf '[core]\nssh_key = %s\n' "$LAUNCH_DIR/sandbox-key" >> "$conf"
+    # No default, so a config that launches must select one.
+    printf '[core]\nharness = claude-code\nssh_key = %s\n' "$LAUNCH_DIR/sandbox-key" >> "$conf"
     HOME="$LAUNCH_DIR/home" PATH="$LAUNCH_DIR:$PATH" \
         SANDBOX_TEST_DOCKER_LOG="$LAUNCH_DIR/docker-log" \
         SANDBOX_TEST_SETTINGS_COPY="$LAUNCH_DIR/settings-copy.json" \
         SSH_AUTH_SOCK= ALLOWED_DOMAINS= \
-        "$REPO_ROOT/bin/dejima" "$@"
+        "${DEJIMA_BIN:-$REPO_ROOT/bin/dejima}" "$@"
 }
 
 run_line() { grep -- ' run --rm ' "$LAUNCH_DIR/docker-log" | tail -1; }
@@ -280,15 +339,9 @@ done
 exit 0
 EOF
     chmod +x "$LAUNCH_DIR/docker"
-    AUDIT_REAL="$REPO_ROOT/bin/dejima-audit"
-    AUDIT_BACKUP="$(mktemp)"
-    cp "$AUDIT_REAL" "$AUDIT_BACKUP"
-    printf '#!/bin/bash\nexit 3\n' > "$AUDIT_REAL"
-    chmod +x "$AUDIT_REAL"
+    shadow_repo
+    shadow_failing_audit
     run launch "$PROJ"
-    cp "$AUDIT_BACKUP" "$AUDIT_REAL"
-    chmod +x "$AUDIT_REAL"
-    rm -f "$AUDIT_BACKUP"
     [ "$status" -eq 5 ]
     rm -rf "$PROJ"
 }
@@ -309,14 +362,9 @@ done
 exit 0
 EOF
     chmod +x "$LAUNCH_DIR/docker"
-    AUDIT_REAL="$REPO_ROOT/bin/dejima-audit"
-    AUDIT_BACKUP="$(mktemp)"
-    cp "$AUDIT_REAL" "$AUDIT_BACKUP"
-    rm -f "$AUDIT_REAL"
+    shadow_repo
+    rm -f "$SHADOW/bin/dejima-audit"
     run launch "$PROJ"
-    cp "$AUDIT_BACKUP" "$AUDIT_REAL"
-    chmod +x "$AUDIT_REAL"
-    rm -f "$AUDIT_BACKUP"
     [ "$status" -eq 5 ]
     rm -rf "$PROJ"
 }
@@ -2450,14 +2498,17 @@ nginx_blocks() {
 }
 
 @test "install.sh links the remote scripts and the notify hook, registers neither, and enables nothing" {
-    grep -q '^link_hook dejima-notify.sh$' "$REPO_ROOT/install.sh"
+    # The hook is linked by the harness plugin now - which file it is, and
+    # where the harness reads it from, is the harness's business - while the
+    # CLIs stay install.sh's, being dejima's own.
+    grep -q '_link_hook dejima-notify.sh' "$REPO_ROOT/bin/harness/claude-code.sh"
     grep -q '^link_bin dejima-attach$' "$REPO_ROOT/install.sh"
     grep -q '^link_bin dejima-ttyd$' "$REPO_ROOT/install.sh"
     grep -q '^link_bin dejima-notify-watch$' "$REPO_ROOT/install.sh"
     # The notify hook is registered per session by bin/dejima, into a
     # throwaway settings copy. Registered here it would fire for every
     # unsandboxed host session too.
-    run grep -nE 'register_hook.*dejima-notify' "$REPO_ROOT/install.sh"
+    run grep -nE '_register .*dejima-notify' "$REPO_ROOT/bin/harness/claude-code.sh"
     [ "$status" -ne 0 ]
     # Standing up a front door is an explicit operator decision. Nothing in
     # the installer may start, enable or even copy a unit. Code only - the

test/runtime.bats

@@ -29,42 +29,31 @@ IMAGE=dejima:latest
 container_name() { echo "dejima-test-$(basename "$BATS_FILE_TMPDIR")"; }
 network_name() { echo "dejima-test-net-$(basename "$BATS_FILE_TMPDIR")"; }
 
-# Every container this file starts must go on network_name(), never the
-# default bridge: dns-proxy.sh forwards allowlisted names to Docker's
-# embedded resolver at 127.0.0.11, which only exists on a user-defined
-# network (compose gives the real sandbox one). On the default bridge the
-# container gets the host's nameservers instead, dnsmasq has nowhere to
-# forward, and startup dies in dns-proxy.sh - see its own check for this.
+# Every container here must go on network_name(), never the default bridge:
+# dns-proxy.sh forwards to Docker's embedded resolver at 127.0.0.11, which
+# only exists on a user-defined network. On the default bridge dnsmasq has
+# nowhere to forward and startup dies in dns-proxy.sh.
 #
-# Readiness is the *final* `sleep infinity` (entrypoint.sh's
-# SANDBOX_TEST_MODE exec), matched exactly. A bare `pgrep -x sleep` also
-# matches the short-lived sleeps of the startup sequence itself
-# (dns-proxy.sh's dig-retry loop, entrypoint.sh's forwarder-bind wait), and
-# a test that proceeds on one of those asserts against a container whose
-# firewall hasn't been installed yet - which silently turns egress
-# assertions into vacuous passes, since everything is still reachable at
-# that point. Waiting for `sleep infinity` means the whole startup sequence
-# (DNS proxy, firewall, ipsets, logging, forwarder bind) has completed.
+# Readiness is the FINAL `sleep infinity` (entrypoint.sh's SANDBOX_TEST_MODE
+# exec), matched exactly. A bare `pgrep -x sleep` also matches the startup
+# sequence's own short sleeps, and proceeding on one of those asserts against
+# a container with no firewall yet - turning egress assertions into vacuous
+# passes, since everything is still reachable.
 #
-# 120 rather than 60: init-firewall.sh's GitHub ranges fetch now retries, and a
-# run that loses two attempts before succeeding spends ~26s of the budget on
-# that call alone (#59) - which on a runner that has just spent three minutes
-# unpacking an image is enough to turn a recovery into a spurious failure. A
-# longer budget only delays reporting a genuinely stuck startup; it cannot hide
-# one, since nothing here passes on a timeout.
-# `docker wait` returns the moment the container exits, but the json-file log
-# driver writes asynchronously - so reading the log right then can return one
-# that stops mid-stream, with everything after some flush boundary missing.
+# 120s rather than 60: the GitHub ranges fetch retries, and losing two
+# attempts costs ~26s (#59), which on a runner that just spent minutes
+# unpacking an image turns a recovery into a spurious failure. A longer
+# budget only delays reporting a stuck startup; nothing here passes on a
+# timeout.
+# `docker wait` returns on exit, but the json-file log driver writes
+# asynchronously, so reading right then can return a log that stops
+# mid-stream. That presented as a container "aborting" partway through
+# init-firewall.sh's domain loop at a different domain each run, and survived
+# three fixes before the giveaway: no message from an ERR trap that cannot be
+# bypassed. Nothing had aborted; the read was short.
 #
-# That is not hypothetical. It presented as a container "aborting" partway
-# through init-firewall.sh's domain loop, at a slightly different domain each
-# run, and survived three separate fixes to that loop before the giveaway: the
-# log ended with no message from an ERR trap that cannot be bypassed and no
-# output from a retry that prints on every path. Nothing had aborted. The read
-# was short.
-#
-# Two identical consecutive reads means the driver has caught up. Deliberately
-# content-neutral - waiting for an expected string would make the wait pass or
+# Two identical consecutive reads means the driver has caught up.
+# Content-neutral on purpose - waiting for an expected string would pass or
 # fail on the very thing the caller is about to assert.
 settled_logs() {
     local container="$1" current previous=""
@@ -408,8 +397,30 @@ dexec_node() {
 }
 
 @test "GitHub is reachable on port 443" {
-    run dexec curl --connect-timeout 5 -s -o /dev/null -w "%{http_code}" https://api.github.com/zen
-    [ "$status" -eq 0 ]
+    # Retried, and only because this is a POSITIVE assertion: more attempts
+    # can only turn a pass into a pass. The unreachability assertions below
+    # are never retried - there, a retry could hide a hole that opens once.
+    #
+    # It failed once on CI with nothing to go on: no curl exit code, no HTTP
+    # status, no way to tell DNS from the ipset from the far end. Everything
+    # else in that run said the firewall was right - the port-22 push grant
+    # reached GitHub, so github-domains held real addresses - which left one
+    # five-second connect against a live third-party service as the only
+    # explanation, and no evidence to confirm it with. Hence the diagnostics
+    # as much as the retries.
+    local attempt code rc=0
+    for attempt in 1 2 3; do
+        code="$(dexec curl --connect-timeout 5 -s -o /dev/null \
+                     -w '%{http_code}' https://api.github.com/zen 2>&1)" && return 0
+        rc=$?
+        sleep 2
+    done
+    echo "api.github.com:443 unreachable after 3 attempts." >&2
+    echo "last curl exit: $rc, last output: '$code'" >&2
+    echo "exit 6 is DNS, 7 or 28 is the route or the firewall; an HTTP code" >&2
+    echo "with a non-zero exit means the connection worked. Check with" >&2
+    echo "'ipset list github-domains' inside the container." >&2
+    false
 }
 
 @test "GitHub is NOT reachable on port 22 (push transport is physically absent)" {
@@ -594,6 +605,73 @@ dexec_node() {
     rm -rf "$src"
 }
 
+# #104: the whole ~/.claude/projects tree used to be mounted read-write, so a
+# session could write another project's transcripts and agent memory. Memory
+# is loaded into context at session start, which makes that a write into a
+# future session's premises rather than its input.
+#
+# Three properties, and the second is the one that decides whether scoping is
+# viable at all: a slug that is absent has to be creatable, because that is
+# what Claude Code does on a first run against a path with no history. If the
+# parent were read-only instead, this is where it would hang the way a
+# blocked ~/.claude.json write did.
+@test "only the mounted project slugs exist in the container, and an unmounted one is still creatable" {
+    local test_container="dejima-test-projects-$(basename "$BATS_FILE_TMPDIR")"
+    local src; src="$(mktemp -d)"
+    mkdir -p "$src/own" "$src/sibling" "$src/fh-own"
+    echo host > "$src/own/existing.jsonl"
+    echo secret > "$src/sibling/other.jsonl"
+    # file-history holds verbatim copies of edited files, so an unmounted
+    # session directory being absent is the whole point of scoping it.
+    echo 'other project source' > "$src/fh-own/aaaa@v1"
+    chmod -R 777 "$src"
+    docker run -d --name "$test_container" \
+        --cap-add=NET_ADMIN --cap-add=NET_RAW \
+        --network "$(network_name)" \
+        -e SANDBOX_TEST_MODE=1 \
+        -e GH_TOKEN="${GH_TOKEN:-}" \
+        -v "$src/own:/home/node/.claude/projects/own" \
+        -v "$src/fh-own:/home/node/.claude/file-history/own-session" \
+        --entrypoint /usr/local/bin/entrypoint.sh \
+        "$IMAGE" >/dev/null
+    wait_ready "$test_container"
+
+    # the sibling project is absent, not merely unwritable
+    run docker exec -u node "$test_container" test -e /home/node/.claude/projects/sibling
+    [ "$status" -ne 0 ]
+    run docker exec -u node "$test_container" ls /home/node/.claude/projects
+    [[ "$output" != *sibling* ]]
+
+    # a slug with no history yet can still be created, and stays in the
+    # container: it is not on the host afterwards
+    run docker exec -u node "$test_container" \
+        bash -c 'mkdir -p /home/node/.claude/projects/fresh && echo x > /home/node/.claude/projects/fresh/a.jsonl'
+    [ "$status" -eq 0 ]
+    [ ! -e "$src/fresh" ]
+
+    # and the mounted slug is the one that does reach the host
+    docker exec -u node "$test_container" \
+        bash -c 'echo written > /home/node/.claude/projects/own/new.jsonl'
+    [ "$(cat "$src/own/new.jsonl")" = written ]
+
+    # Same three properties for file-history. The middle one is what the
+    # image's pre-created directory buys: without it Docker creates
+    # file-history/ root-owned while setting up the mount nested in it, and
+    # the session cannot record an undo entry for its own id at all.
+    run docker exec -u node "$test_container" ls /home/node/.claude/file-history
+    [[ "$output" != *other-session* ]]
+    run docker exec -u node "$test_container" \
+        bash -c 'mkdir -p /home/node/.claude/file-history/fresh-session && echo x > /home/node/.claude/file-history/fresh-session/b@v1'
+    [ "$status" -eq 0 ]
+    [ ! -e "$src/fresh-session" ]
+    docker exec -u node "$test_container" \
+        bash -c 'echo written > /home/node/.claude/file-history/own-session/c@v1'
+    [ "$(cat "$src/fh-own/c@v1")" = written ]
+
+    docker rm -f "$test_container" >/dev/null 2>&1 || true
+    rm -rf "$src"
+}
+
 @test "a staged statusline.sh is copied in root-owned, and replacing it never reaches the host" {
     # Companion to the overlay-dirs test above, but for the config
     # bin/dejima stages at /etc/dejima-claude for entrypoint.sh to copy into

test/static.bats

@@ -449,16 +449,98 @@ EOF
     # Directories, which is what makes them safe: a directory bind mount
     # can't detach the way a file one does, because a rename replaces an
     # entry *within* the mounted directory rather than the mount point.
-    # projects/ is what host-side `claude --resume` reads; the rest is the
-    # edit-undo record and the task/plan state skills write.
-    grep -qE '^\s*-\s*\$\{HOME\}/\.claude/projects:\$\{HOME\}/\.claude/projects\s*$' "$SANDBOX_DIR/docker-compose.yml"
-    grep -qE '^\s*-\s*\$\{HOME\}/\.claude/file-history:\$\{HOME\}/\.claude/file-history\s*$' "$SANDBOX_DIR/docker-compose.yml"
+    # projects/ and file-history/ are not in this list: both are mounted an
+    # entry at a time by bin/dejima, see the tests below. The rest is the
+    # task/plan state skills write and the ~/.claude.json backups.
     grep -qE '^\s*-\s*\$\{HOME\}/\.claude/backups:\$\{HOME\}/\.claude/backups\s*$' "$SANDBOX_DIR/docker-compose.yml"
     grep -qE '^\s*-\s*\$\{HOME\}/\.claude/tasks:\$\{HOME\}/\.claude/tasks\s*$' "$SANDBOX_DIR/docker-compose.yml"
     grep -qE '^\s*-\s*\$\{HOME\}/\.claude/plans:\$\{HOME\}/\.claude/plans\s*$' "$SANDBOX_DIR/docker-compose.yml"
     grep -qE '^\s*-\s*\$\{HOME\}/\.claude/sessions:\$\{HOME\}/\.claude/sessions\s*$' "$SANDBOX_DIR/docker-compose.yml"
 }
 
+# The whole tree carries every project's transcripts and agent memory, and
+# memory is loaded into context at session start - so write access to it is
+# enough to plant text a future session in another project reads as
+# background (#104). This is the assertion that the tree itself never comes
+# back; the per-slug mounts that replace it are covered below.
+# The service still refers to the static keys; only the volume Docker
+# actually creates varies, so a hardcoded name here would silently put every
+# project back on one shared set.
+@test "the nvim volumes take their names from the environment" {
+    grep -qE '^\s*name:\s*\$\{SANDBOX_NVIM_DATA:-' "$SANDBOX_DIR/docker-compose.yml"
+    grep -qE '^\s*name:\s*\$\{SANDBOX_NVIM_STATE:-' "$SANDBOX_DIR/docker-compose.yml"
+    grep -qE '^\s*name:\s*\$\{SANDBOX_NVIM_CACHE:-' "$SANDBOX_DIR/docker-compose.yml"
+    grep -q 'export SANDBOX_NVIM_DATA=' "$REPO_ROOT/bin/dejima"
+}
+
+@test "docker-compose.yml never mounts the whole ~/.claude/projects tree" {
+    run grep -nE '^\s*-\s*\$\{HOME\}/\.claude/projects:' "$SANDBOX_DIR/docker-compose.yml"
+    [ "$status" -ne 0 ] || {
+        echo "the projects tree is mounted whole again:" >&2
+        echo "$output" >&2
+        false
+    }
+}
+
+@test "the claude-code plugin mounts projects/ one slug at a time" {
+    # Moved out of bin/dejima into the harness plugin: which directories a
+    # harness keeps, and what decides an entry belongs to this workspace, is
+    # a fact about the harness. The property is unchanged - the tree is
+    # every project's transcripts and memory, and memory is read into
+    # context at session start (#104).
+    local p="$REPO_ROOT/bin/harness/claude-code.sh"
+    grep -q 'workspace_slugs_sorted "$ws"' "$p"
+    grep -q 'emit_mount "$projects/$slug"' "$p"
+}
+
+# Docker creates the parents of a mount point root-owned when they do not
+# already exist. projects/ is now such a parent, so without this the session
+# could not create a slug of its own - which is exactly what it must do when
+# the workspace has no history yet.
+@test "the image pre-creates ~/.claude/projects so the slug mounts have a writable parent" {
+    grep -qE '^RUN mkdir -p .*\$SANDBOX_HOME/\.claude/projects' "$SANDBOX_DIR/Dockerfile"
+}
+
+# Each file-history/<session>/<hash>@v1 is a verbatim copy of an edited file,
+# so mounting the tree hands a session every project's source outright -
+# worse than the transcripts, which are a record about the work rather than
+# the work. Writable too, so tampering with another project's undo record
+# puts chosen content into a later restore there.
+@test "docker-compose.yml never mounts the whole ~/.claude/file-history tree" {
+    run grep -nE '^\s*-\s*\$\{HOME\}/\.claude/file-history:' "$SANDBOX_DIR/docker-compose.yml"
+    [ "$status" -ne 0 ] || {
+        echo "the file-history tree is mounted whole again:" >&2
+        echo "$output" >&2
+        false
+    }
+}
+
+@test "the claude-code plugin mounts file-history one session at a time" {
+    local p="$REPO_ROOT/bin/harness/claude-code.sh"
+    grep -q 'workspace_session_ids "$projects" "$ws"' "$p"
+    grep -q 'emit_mount "$history/$sid"' "$p"
+}
+
+@test "the image pre-creates ~/.claude/file-history so the session mounts have a writable parent" {
+    grep -qE '\$SANDBOX_HOME/\.claude/file-history' "$SANDBOX_DIR/Dockerfile"
+}
+
+# One definition of "this workspace's slugs", shared by the mount and the
+# audit. If they drift, a slug the mount exposes but the audit calls foreign
+# fires a finding every session, and a slug the audit calls its own but the
+# mount does not expose is a hole nothing reports.
+@test "bin/dejima and bin/dejima-audit share one slug definition" {
+    grep -q 'source "$SCRIPT_DIR/dejima-slug-lib.sh"' "$REPO_ROOT/bin/dejima"
+    grep -q 'dejima-slug-lib.sh"$' "$REPO_ROOT/bin/dejima-audit"
+    # and neither keeps a private copy of the sed scheme
+    run grep -n "s/\[^a-zA-Z0-9\]/-/g" "$REPO_ROOT/bin/dejima" "$REPO_ROOT/bin/dejima-audit"
+    [ "$status" -ne 0 ] || {
+        echo "an inline copy of the slug scheme is back:" >&2
+        echo "$output" >&2
+        false
+    }
+}
+
 @test "~/.claude config files are staged for copying, not mounted at their real paths" {
     # A copy can't be detached by a host-side rename at all. entrypoint.sh
     # installs them root-owned and read-only; see its comment and the mount
@@ -617,8 +699,13 @@ EOF
     # Which only takes effect if the flag is actually passed. It goes in the
     # wrapper rather than the entrypoint so it also covers invocations the
     # entrypoint never sees (an interactive shell, claudecode.nvim).
-    grep -qE -- '--settings /etc/dejima-settings\.json' "$SANDBOX_DIR/claude-wrapper.sh"
-    grep -qE -- '"\$\{settings\[@\]\}"' "$SANDBOX_DIR/claude-wrapper.sh"
+    # The flag itself is the harness's now, not the wrapper's: the wrapper
+    # is shared and a --settings it hardcoded would be handed to a harness
+    # that layers settings some other way.
+    grep -qE -- 'settings_args\s*=\s*--settings /etc/dejima-settings\.json' \
+        "$REPO_ROOT/harnesses/claude-code/harness.conf"
+    grep -q 'SANDBOX_HARNESS_SETTINGS' "$SANDBOX_DIR/harness-wrapper.sh"
+    grep -qE -- '\$\{settings\[@\]\+"\$\{settings\[@\]\}"\}' "$SANDBOX_DIR/harness-wrapper.sh"
 }
 
 @test "bin/dejima pre-creates the ~/.claude config files so neither is mounted as a directory" {
@@ -1089,9 +1176,13 @@ cli_payload() { cat "$BATS_TEST_TMPDIR/cli/payload"; }
     [ "$status" -eq 0 ]
 }
 
-@test "install.sh links and registers the merge guard" {
-    grep -qx 'link_hook merge-guard.sh' "$REPO_ROOT/install.sh"
-    grep -q 'register_hook PreToolUse Bash merge-guard.sh' "$REPO_ROOT/install.sh"
+@test "the merge guard is linked and registered" {
+    # In the harness plugin now: which file a guard is and where a harness
+    # reads it from is the harness's, while whether it is installed at all
+    # is gated on the manifest declaring the guard.
+    local p="$REPO_ROOT/bin/harness/claude-code.sh"
+    grep -q 'harness_has_guard merge-guard && _link_hook merge-guard.sh' "$p"
+    grep -q '_register PreToolUse Bash merge-guard.sh' "$p"
 }
 
 # Runs install.sh against a throwaway HOME, with a stub `docker` on PATH so
@@ -1341,15 +1432,46 @@ registered_commands() {
     audit_line=$(grep -n 'dejima-audit" compare' "$REPO_ROOT/bin/dejima" | head -1 | cut -d: -f1)
     trap_line=$(grep -n '^trap cleanup EXIT' "$REPO_ROOT/bin/dejima" | head -1 | cut -d: -f1)
     marker_line=$(grep -n 'could not remove the marker directory' "$REPO_ROOT/bin/dejima" | head -1 | cut -d: -f1)
+    epitaph_line=$(grep -n 'dejima-epitaph" save' "$REPO_ROOT/bin/dejima" | head -1 | cut -d: -f1)
     [ -n "$marker_line" ]
+    [ -n "$epitaph_line" ]
     [ "$audit_line" -lt "$trap_line" ]
     # Not just "inside cleanup()" but after the one line in it a session can
     # deliberately provoke to print - the actual last-writer invariant.
     [ "$marker_line" -lt "$audit_line" ]
+    # Same invariant, for the other task in the trap that prints: epitaph
+    # announces every summary it keeps, so it goes before the verdict.
+    [ "$epitaph_line" -lt "$audit_line" ]
     run grep -c 'could not remove the marker directory' "$REPO_ROOT/bin/dejima"
     [ "$output" -eq 1 ]
 }
 
+@test "cleanup says what it is doing before each task that takes a while" {
+    # A silent terminal between the container exiting and the prompt
+    # returning reads as a hung one, and the operator's answer to that is
+    # Ctrl-C - during the audit, which is the worst moment for it.
+    local d="$REPO_ROOT/bin/dejima" start s1 epi s2 aud
+    # Absolute line numbers, so each announcement is checked to sit with the
+    # task it names rather than merely somewhere in the same function. The
+    # definition of exit_step() itself is above cleanup() and so is skipped.
+    start=$(grep -n '^cleanup() {' "$d" | cut -d: -f1)
+    epi=$(grep -n 'dejima-epitaph" save' "$d" | head -1 | cut -d: -f1)
+    aud=$(grep -n 'dejima-audit" compare' "$d" | head -1 | cut -d: -f1)
+    s1=$(awk -v s="$start" 'NR>s && /exit_step /{print NR; exit}' "$d")
+    s2=$(awk -v s="$epi" 'NR>s && /exit_step /{print NR; exit}' "$d")
+    [ -n "$s1" ] && [ -n "$s2" ] || { echo "no announcement before a task"; return 1; }
+    [ "$s1" -lt "$epi" ] || { echo "epitaph runs unannounced"; return 1; }
+    [ "$s2" -lt "$aud" ] || { echo "the audit runs unannounced"; return 1; }
+    # Gated on the snapshot, the same condition `compare` uses to decide it
+    # has work: an exit before the session started is instant, and its last
+    # line is the error that caused it - which announcements would bury.
+    run sed -n '/^exit_step() {/,/^}/p' "$REPO_ROOT/bin/dejima"
+    [ -n "$output" ] || { echo "exit_step is not defined"; return 1; }
+    grep -q 'SANDBOX_AUDIT_STATE' <<< "$output"
+    # A nonzero status from anything in an EXIT trap replaces the script's.
+    grep -q 'return 0' <<< "$output"
+}
+
 @test "credential proxy is on by default, gated only by credentials.proxy" {
     grep -q 'config_get_bool credentials proxy' "$REPO_ROOT/bin/dejima"
     grep -Pq 'credentials\.proxy\tbool\ton' "$REPO_ROOT/bin/dejima-config-lib.sh"
@@ -1503,9 +1625,12 @@ registered_commands() {
     # coverage (snapshot/compare
     # against real fixtures, including the symlink-swap exploit) lives in
     # test/audit.bats; this just confirms the wiring exists.
-    grep -q 'STATUSLINE_PATH=' "$REPO_ROOT/bin/dejima-audit"
-    grep -q 'statusline_snapshot' "$REPO_ROOT/bin/dejima-audit"
-    grep -q 'HOST-CODE-EXECUTION' "$REPO_ROOT/bin/dejima-audit"
+    # The path and its label are the harness plugin's now; the mechanism
+    # that snapshots and reports them is still the audit's.
+    grep -q 'host_exec_snapshot' "$REPO_ROOT/bin/dejima-audit"
+    grep -q 'host_exec_load' "$REPO_ROOT/bin/dejima-audit"
+    grep -q 'statusline_path' "$REPO_ROOT/bin/harness/claude-code.sh"
+    grep -q 'HOST-CODE-EXECUTION' "$REPO_ROOT/bin/harness/claude-code.sh"
 }
 
 @test "Dockerfile installs the sandbox clipboard fallback into nvim's VIMRUNTIME plugin dir" {
@@ -1674,6 +1799,13 @@ this line is not json'
     [ "$status" -ne 0 ]
     run grep -nE '^\s*extra_mounts\+=.*(PROJECT_STORE|config_root)' "$REPO_ROOT/bin/dejima"
     [ "$status" -ne 0 ]
+    # Both greps are literal, so neither can see a mount whose path is
+    # computed at run time - and two of those reached the store (#15): a
+    # [mounts] readonly entry naming it, and a ~/.claude symlink pointing
+    # into it. What refuses them is asserted by running the launcher, in
+    # config.bats ("refused when it names the approval store"). This test
+    # is the cheap half and does not stand in for that one.
+    grep -q '_refuse_guarded_mount' "$REPO_ROOT/bin/dejima"
 }
 
 @test "sandbox-context escapes a declared domain's dots before matching the servers file" {
@@ -1905,12 +2037,15 @@ EOF
     grep -q 'not a verdict on who caused them' "$REPO_ROOT/README.md"
 }
 
-@test "docs/security.md documents the ~/.claude/projects cross-session false-positive case" {
-    # Presence-only grep, same caveat as above: confirms the specific
-    # ~/.claude/projects wording landed, not that the surrounding
-    # reasoning is sound.
-    grep -q 'CROSS-PROJECT. finding' "$REPO_ROOT/docs/security.md"
-    grep -q 'directory shared by' "$REPO_ROOT/docs/security.md"
+# Was: that the cross-session false-positive case is documented. The finding
+# itself is gone - a session can no longer write another project's slug, so
+# every one it could still see belonged to some other session - and what
+# needs documenting now is its absence, so that removing detection stays a
+# stated decision rather than something a reader has to infer from silence.
+@test "docs/security.md says why there is no cross-project finding" {
+    grep -q 'no longer a .CROSS-PROJECT. finding' "$REPO_ROOT/docs/security.md"
+    run grep -n 'CROSS-PROJECT history/memory write' "$REPO_ROOT/bin/dejima-audit"
+    [ "$status" -ne 0 ] || { echo "the finding is back but the doc says it is gone" >&2; false; }
 }
 
 # The CI disclosure is the one part of issue #72 that helps a reader whatever
@@ -1924,6 +2059,46 @@ EOF
     done
 }
 
+@test "core never reads a harness's config by a literal harness name" {
+    # `keys` is declared per harness, so a fixed [harness.<name>] at a call
+    # site makes it a declaration only one harness can ever be read through:
+    # a second harness is selected, is validated, and is then handed the
+    # first one's paths with nothing erroring anywhere. A silent wrong
+    # answer, so it gets a test rather than a convention.
+    run grep -rn 'config_get["'"'"' ]*harness\.[a-z]' "$REPO_ROOT/bin"
+    [ "$status" -ne 0 ] || { echo "use harness_config_get instead:"; echo "$output"; return 1; }
+    grep -q '^harness_config_get()' "$REPO_ROOT/bin/dejima-provider-lib.sh"
+}
+
+@test "the harness guide states the bar a harness has to clear" {
+    # Support is per harness and deliberate, at the depth Claude Code has.
+    # That only holds if the requirements are written down: without them
+    # each new harness is argued about rather than measured, and the first
+    # one that cannot host a guard gets waved through.
+    local sec r
+    grep -q '^## Adding a new harness' "$REPO_ROOT/docs/development.md"
+    sec="$(sed -n '/^### What a harness must be able to do/,/^### What the manifest/p' \
+        "$REPO_ROOT/docs/development.md")"
+    [ -n "$sec" ] || { echo "the capability bar is gone"; return 1; }
+    for r in R1 R2 R3 R4 R5; do
+        grep -q "\*\*$r\*\*" <<< "$sec" || { echo "no $r in the bar"; return 1; }
+    done
+    # Unattended execution is the premise every other control is justified
+    # against, so it stays stated rather than assumed.
+    grep -qi 'without confirming each one' <<< "$sec"
+}
+
+@test "a harness dejima declines is declined against a stated requirement" {
+    # The reason has to point at a rule. Otherwise the bar is decoration and
+    # the next harness is decided on how much someone wants it.
+    local sec
+    sec="$(sed -n '/^### Worked example/,/^### What the manifest/p' \
+        "$REPO_ROOT/docs/development.md")"
+    [ -n "$sec" ] || { echo "no worked example of a refusal"; return 1; }
+    grep -q 'R2' <<< "$sec" \
+        || { echo "the refusal names no requirement it failed"; return 1; }
+}
+
 @test "docs/ci.md exists and is linked from the README and docs/security.md" {
     [ -f "$REPO_ROOT/docs/ci.md" ]
     grep -q 'docs/ci\.md' "$REPO_ROOT/README.md"
@@ -1943,6 +2118,81 @@ EOF
 # likely to skip: a project with no CI is exposed too, because a branch adds
 # the first workflow. If that goes missing, the document reads as being about
 # people who already run CI, which is the smaller group.
+# Who wrote the prose. The repository argues at length in a voice that is not
+# its maintainer's, and saying so is the difference between disclosure and
+# letting a reader assume. Pinned because it is the kind of line that goes
+# quietly in a tidy-up, and its absence is not visible from anything else.
+@test "the README says the prose is written by agents" {
+    grep -q 'A note from the developer' "$REPO_ROOT/README.md"
+    grep -q 'the words are written rapidly' "$REPO_ROOT/README.md"
+    # the contrast that makes both claims credible: ATTRIBUTION.md says the
+    # artwork is the other way round, and the README points at it
+    grep -q 'no AI tools were used to create it' "$REPO_ROOT/README.md"
+    grep -q 'No AI tools were used' "$REPO_ROOT/ATTRIBUTION.md"
+}
+
+# The maturity disclaimer, pinned in all three places a reader can land. A
+# project that argues at length for its own security properties is exactly the
+# one that has to say those properties are unreviewed - and it is the first
+# thing that gets quietly dropped as a project starts to feel established.
+@test "every entry point says this is unreviewed early development" {
+    local f
+    for f in README.md SECURITY.md docs/security.md; do
+        grep -q 'reviewed by anyone else\|outside review' "$REPO_ROOT/$f" || {
+            echo "$f does not say the work is unreviewed" >&2
+            false
+        }
+    done
+    grep -q 'Not ready to be depended on' "$REPO_ROOT/README.md"
+    grep -q 'not proven software' "$REPO_ROOT/SECURITY.md"
+}
+
+# The GitHub push default was documented as mechanism only - port 443, port 22
+# unreachable - and a reader who needed the reason inferred the wrong one: that
+# it is a containment measure against a hostile session. It is not. The
+# repositories on the other side are other people's, and the failure it
+# prevents takes no compromise, only an agent being helpful at a stranger's
+# expense. Pinned in both places a reader lands, because getting this wrong
+# leads someone to turn the flag on believing it costs them nothing.
+@test "the GitHub push default is documented with its reason, not just its mechanism" {
+    # Matched on a fragment rather than the whole sentence: it wraps
+    # differently in each file, and a line-based grep cannot span that.
+    local f
+    for f in README.md docs/security.md; do
+        grep -q 'not being used to push' "$REPO_ROOT/$f" || {
+            echo "$f gives the mechanism without the principle" >&2
+            false
+        }
+    done
+    # the eager-agent case, which is what makes the principle concrete - and
+    # that it is not softened into a well-meaning near-miss. The unsolicited
+    # pull request is the thing being prevented, not a good deed with bad
+    # timing; docs/responsible-use.md is where the cost to maintainers is set
+    # out, and this must not read as disagreeing with it.
+    grep -q 'an eager agent, not a compromised one' "$REPO_ROOT/docs/security.md"
+    grep -q 'does not make it harmless' "$REPO_ROOT/docs/security.md"
+    # and the ordering a reader got backwards: CI follows from the principle
+    # rather than being the reason for it
+    grep -q 'not the$\|reason for it' "$REPO_ROOT/docs/security.md"
+    grep -q 'dont-point-it-at-repositories-you-dont-own' "$REPO_ROOT/docs/security.md"
+}
+
+# The branch convention is only a convention - nothing refuses a pull request
+# aimed at main - so the three places an agent or a contributor looks have to
+# agree, or the one that is stale becomes the instruction someone follows.
+@test "the development-branch convention is stated where it is looked for" {
+    local f
+    for f in .dejima.md CONTRIBUTING.md docs/development.md; do
+        grep -q 'development' "$REPO_ROOT/$f" || {
+            echo "$f does not mention the development branch" >&2
+            false
+        }
+    done
+    # and the one thing that is easy to get wrong: gitea-pr's default base
+    grep -q -- '--base development' "$REPO_ROOT/.dejima.md"
+    grep -q -- '--base development' "$REPO_ROOT/docs/development.md"
+}
+
 @test "docs/ci.md says a project with no CI is exposed too" {
     grep -q 'applies to projects with no CI at all' "$REPO_ROOT/docs/ci.md"
     grep -q 'no workflows is not therefore safe' "$REPO_ROOT/docs/ci.md"
@@ -2009,6 +2259,31 @@ EOF
     [ "$missing" -eq 0 ]
 }
 
+# Other repositories in the organisation link to these two by absolute URL
+# rather than keeping copies, so the terms cannot drift into two divergent
+# statements. A rename here is a 404 in a repository whose tests cannot see
+# this one.
+@test "the org-wide policy files other repositories link to keep their names" {
+    [ -f "$REPO_ROOT/SECURITY.md" ]
+    [ -f "$REPO_ROOT/ATTRIBUTION.md" ]
+}
+
+# The wordmark is referenced by a fixed filename on purpose - see the comment
+# above the tag - which means a rename is caught here or not at all.
+# Markdown image syntax as well as <img>, and svg as well as png: the badge
+# is referenced the markdown way and is the case this exists for.
+@test "every image the README points at is committed" {
+    local src missing=0
+    for src in $(grep -ohE 'src="[^"]+\.(png|svg)"|\]\([^)]+\.(png|svg)\)' "$REPO_ROOT/README.md" \
+        | sed 's/^src="//; s/"$//; s/^](//; s/)$//'); do
+        if [ ! -f "$REPO_ROOT/$src" ]; then
+            echo "README points at $src, which does not exist" >&2
+            missing=1
+        fi
+    done
+    [ "$missing" -eq 0 ]
+}
+
 # --- the labels this repo's own workflow asks for ---
 #
 # Whether a job container gets the Docker socket is settled by the runner's
@@ -2125,7 +2400,7 @@ EOF
 # This catches address- and hostname-shaped leaks only. A personal domain
 # that looks like any other public one is not mechanically distinguishable
 # and still rests on review.
-SCANNED_PATHS=(README.md docs docker bin forges workflows harnesses hooks install.sh systemd .gitea)
+SCANNED_PATHS=(README.md SECURITY.md ATTRIBUTION.md docs docker bin forges workflows harnesses hooks install.sh systemd .gitea)
 
 @test "no private or link-local address is used as a documentation example" {
     run grep -rnIE '\b(10|127\.0\.0\.[2-9]|169\.254|192\.168|172\.(1[6-9]|2[0-9]|3[01]))\.[0-9]{1,3}\.[0-9]{1,3}\b' \
@@ -2198,6 +2473,7 @@ EOF
     # past validation.
     cat > "$DEJIMA_T/config/config" <<EOF
 [core]
+harness = claude-code
 ssh_key = $DEJIMA_T/sandbox-key
 forge = none
 workflow = local-only
@@ -2468,39 +2744,26 @@ q
 }
 
 @test "core names no forge" {
-    # The single assertion that keeps this separation from decaying: without
-    # it, the first convenient edit puts a forge name back into core. A
-    # whole-file skip or a substring loose enough to stop catching a real
-    # prose leak would both defeat the point, so each file gets its own
-    # narrow, justified exclusion for the specific lines that are not a
-    # forge coupling - everything else in scope must still match nothing.
+    # What keeps core free of forge names. Each file gets a narrow, justified
+    # exclusion rather than a whole-file skip; everything else must match
+    # nothing. Two scans, different subjects:
     #
-    # Two scans, because they have different subjects:
+    # (1) `gitea` and `github.com` as case-insensitive substrings, which
+    #     catches a PROSE leak ("push to the Gitea remote") as much as a
+    #     coupling.
+    # (2) Every name `provider_list forge` reports, in literal position -
+    #     derived from the shipped directories, so a new forge is covered the
+    #     day it lands. Scan 1 missed exactly that: bare `github` matches
+    #     neither alternative, so core could compare a forge name with this
+    #     test green.
     #
-    # (1) `gitea` and `github.com` as substrings, case-insensitively. This is
-    #     the original scan, and it is what catches a PROSE leak ("push to
-    #     the Gitea remote") as much as a coupling.
-    #
-    # (2) Every name `provider_list forge` reports, in literal position.
-    #     Derived from the shipped directories rather than written out here,
-    #     so a forge is covered the day it lands - which is exactly what
-    #     scan 1 missed: `github` became a shipped FORGE name, and a bare
-    #     `github` matches neither of scan 1's alternatives, so core could
-    #     (and did) compare a forge name to a literal with this test green.
-    #
-    #     "Literal position" means quoted, backticked, bracketed as a config
-    #     section, prefixed `forge.`, under `forges/`, or a `case` arm -
-    #     every shape a lookup or comparison actually takes here. It is not a
-    #     bare-word scan because two shipped names cannot survive one: `none`
-    #     is an ordinary English word this repo's prose uses constantly, and
-    #     `github` is also the name of a CORE feature that predates the forge
-    #     axis (GitHub read-only: `[github] token_file`, the `github.com`
-    #     credential helper, `--allow-github-push`). Operator-facing remedy
-    #     text naming a config VALUE (`set [core] forge = none`, `or [core]
-    #     forge = github`) is deliberately outside these shapes too: a rule
-    #     that refuses a combination has to be able to say which one works,
-    #     and "pick a different forge" without saying which is a worse
-    #     message, not a purer one.
+    #     "Literal position" means quoted, backticked, a config section,
+    #     `forge.`-prefixed, under `forges/`, or a case arm - not a bare-word
+    #     scan, because `none` is an ordinary English word and `github` also
+    #     names a core feature predating the forge axis (`[github]
+    #     token_file`, `--allow-github-push`). Remedy text naming a config
+    #     VALUE is outside these shapes too: a rule that refuses a
+    #     combination has to say which one works.
     local f matches fail=0 name literals="" body
     while IFS= read -r name; do
         [ -n "$name" ] || continue
@@ -2634,6 +2897,27 @@ q
     grep -q 'DEJIMA_CI_TIER: privileged' "$wf"
 }
 
+# test/remote.bats used to stand in a failing exit-time audit by overwriting
+# `$REPO_ROOT/bin/dejima-audit` and restoring it afterwards. The restore does
+# not happen if the run is killed partway, and two runs of that file at once
+# restore each other's stub, so a suite could pass while leaving the real
+# audit disabled - and a file watcher looking at the tree during the window
+# sees a security tool replaced by a two-line `exit 3`, which reads exactly
+# like an attack.
+#
+# Tests get a throwaway copy to mutate instead. Redirects and removals are
+# what actually went wrong, so those are what this pins; a test that finds
+# some third way to write into the checkout is not covered.
+@test "no test suite writes into the repository it is testing" {
+    run grep -rnE '(>>?|rm +(-[a-zA-Z]+ +)*)[[:space:]]*"?\$\{?(REPO_ROOT|BATS_TEST_DIRNAME)' \
+        "$REPO_ROOT"/test/*.bats
+    [ "$status" -ne 0 ] || {
+        echo "a test writes into the checkout; give it a copy instead:" >&2
+        echo "$output" >&2
+        false
+    }
+}
+
 @test "the agent briefing teaches the project scope, not just global" {
     # The observed failure: an agent asked for a project-specific domain to be
     # added container-wide and globally, because --project was never named.
@@ -2693,3 +2977,192 @@ q
     run grep -n 'scope-flag) form\|scope-flag form' "$REPO_ROOT/README.md"
     [ "$status" -ne 0 ]
 }
+
+# --- harness plugins: a plugin computes, core enforces ---
+
+@test "no harness plugin touches the mount list directly" {
+    # A plugin that appended to extra_mounts itself would skip the
+    # guarded-store refusal emit_mount runs at the point of declaration -
+    # and would skip it differently in each harness, which is how one
+    # security model becomes one per harness. The plugins are trusted code,
+    # so this is held at review time rather than by isolating them.
+    run grep -rn 'extra_mounts' "$REPO_ROOT/bin/harness"
+    [ "$status" -ne 0 ] || { echo "a plugin bypasses emit_mount:"; echo "$output"; return 1; }
+}
+
+@test "a harness plugin defines only hooks core actually calls" {
+    # A function named something core never calls looks wired up and is
+    # dead. Worse, a near-miss of a real hook name (harness_state_mount for
+    # harness_state_mounts) reads as implemented and silently does nothing.
+    local hooks f fn
+    hooks="$(sed -n 's/^DEJIMA_HARNESS_HOOKS="\(.*\)"$/\1/p' \
+        "$REPO_ROOT/bin/dejima-harness-lib.sh")"
+    [ -n "$hooks" ] || { echo "the hook list is gone"; return 1; }
+    for f in "$REPO_ROOT"/bin/harness/*.sh; do
+        [ -f "$f" ] || continue
+        while IFS= read -r fn; do
+            [ -n "$fn" ] || continue
+            case " $hooks " in
+                *" $fn "*) ;;
+                *) echo "$(basename "$f") defines $fn, which core never calls"; return 1 ;;
+            esac
+        done < <(grep -oE '^[a-zA-Z_][a-zA-Z0-9_]*\(\)' "$f" | sed 's/()$//')
+    done
+}
+
+@test "every shipped harness has both a manifest and a plugin" {
+    # Support is per harness and deliberate. A manifest with no plugin is a
+    # half-finished integration that would present as a supported harness.
+    local d name
+    for d in "$REPO_ROOT"/harnesses/*/; do
+        [ -d "$d" ] || continue
+        name="$(basename "$d")"
+        [ -f "$d/harness.conf" ] || { echo "$name has no manifest"; return 1; }
+        [ -f "$REPO_ROOT/bin/harness/$name.sh" ] \
+            || { echo "$name has a manifest but no plugin"; return 1; }
+    done
+    for d in "$REPO_ROOT"/bin/harness/*.sh; do
+        [ -f "$d" ] || continue
+        name="$(basename "$d" .sh)"
+        [ -f "$REPO_ROOT/harnesses/$name/harness.conf" ] \
+            || { echo "$name has a plugin but no manifest"; return 1; }
+    done
+}
+
+@test "harness plugins are inside the trusted computing base the audit watches" {
+    # They run on the host, in the launch path. Under harnesses/ they would
+    # be host-executed code a session could edit in dejima's own checkout
+    # without raising a CRITICAL, because tcb_snapshot covers install.sh,
+    # bin/, hooks/, docker/ and test/ - and not the provider directories.
+    [ -d "$REPO_ROOT/bin/harness" ] || { echo "plugins are not under bin/"; return 1; }
+    run grep -rn 'harness_state_mounts\|harness_plugin_load' "$REPO_ROOT/harnesses"
+    [ "$status" -ne 0 ] || { echo "plugin code under harnesses/:"; echo "$output"; return 1; }
+    sed -n '/^tcb_snapshot()/,/^}/p' "$REPO_ROOT/bin/dejima-audit" | grep -q 'for dir in bin '
+}
+
+# --- the wrapper is the harness's, not Claude Code's ---
+
+@test "the wrapper names no harness, and the image installs it under a neutral name" {
+    # It is shadowed into place under the harness's own command by
+    # entrypoint.sh, because the image is shared across harnesses and the
+    # selection is per launch. A COPY to a fixed name would bake one
+    # harness's command into every image.
+    local w="$SANDBOX_DIR/harness-wrapper.sh"
+    [ -f "$w" ] || { echo "the generic wrapper is gone"; return 1; }
+    run grep -niE 'claude|anthropic' "$w"
+    [ "$status" -ne 0 ] || { echo "the shared wrapper names a harness:"; echo "$output"; return 1; }
+    grep -q 'COPY docker/harness-wrapper.sh /usr/local/bin/dejima-harness-wrapper' \
+        "$SANDBOX_DIR/Dockerfile"
+    run grep -n 'COPY docker/claude-wrapper' "$SANDBOX_DIR/Dockerfile"
+    [ "$status" -ne 0 ] || { echo "the old fixed-name copy is back"; return 1; }
+}
+
+@test "the entrypoint refuses a harness command name that is not a bare name" {
+    # It becomes a path under /usr/local/bin. A value with a slash would
+    # write the symlink somewhere else entirely, and this one arrives from
+    # the host as an environment variable.
+    sed -n '/SANDBOX_HARNESS_COMMAND:-/,/^fi/p' "$SANDBOX_DIR/entrypoint.sh" \
+        | grep -q '\*/\*'
+}
+
+@test "the no-confirm flags come from the harness manifest" {
+    # R1 of the capability bar. The flag was hardcoded in the wrapper, which
+    # made unattended operation a property of the image rather than of the
+    # selected harness.
+    grep -qE -- 'no_confirm_args\s*=\s*--dangerously-skip-permissions' \
+        "$REPO_ROOT/harnesses/claude-code/harness.conf"
+    grep -q 'no_confirm_args' "$REPO_ROOT/bin/dejima"
+    # And a harness that names none is refused rather than run asking for
+    # confirmation on every tool call, which presents as a hang.
+    grep -q 'declares no no_confirm_args' "$REPO_ROOT/bin/dejima"
+}
+
+@test "core's schema holds no harness's command as a default" {
+    # [core] command defaulted to the literal `claude`, so core carried one
+    # harness's command name while the manifest's own `command` key was read
+    # by nothing.
+    run grep -nE '^core\.command\s+identifier\s+claude' "$REPO_ROOT/bin/dejima-config-lib.sh"
+    [ "$status" -ne 0 ] || { echo "core still defaults to a harness command"; return 1; }
+    grep -qE '^core\.command\s+\?identifier' "$REPO_ROOT/bin/dejima-config-lib.sh"
+    grep -q 'SANDBOX_CMD="$SANDBOX_HARNESS_COMMAND"' "$REPO_ROOT/bin/dejima"
+}
+
+# --- guards are declared, not assumed ---
+
+@test "every shipped harness declares which guards it can host" {
+    # An absent `guards` reads as "hosts none", which for a harness that
+    # hosts all four would tell the operator at every launch that four
+    # protections are missing. Silence is not a valid declaration here.
+    local d name g known
+    known="$(sed -n 's/^DEJIMA_KNOWN_GUARDS="\(.*\)"$/\1/p' "$REPO_ROOT/bin/dejima-provider-lib.sh")"
+    [ -n "$known" ] || { echo "the known-guard set is gone"; return 1; }
+    for d in "$REPO_ROOT"/harnesses/*/; do
+        [ -f "$d/harness.conf" ] || continue
+        name="$(basename "$d")"
+        grep -qE '^\s*guards\s*=' "$d/harness.conf" \
+            || { echo "$name declares no guards"; return 1; }
+        for g in $(sed -n 's/^\s*guards\s*=\s*//p' "$d/harness.conf"); do
+            case " $known " in
+                *" $g "*) ;;
+                *) echo "$name declares unknown guard: $g"; return 1 ;;
+            esac
+        done
+    done
+}
+
+@test "the guard set core knows lives with the validation that rejects a name" {
+    # It was in dejima-harness-lib.sh while provider_validate_selection read
+    # it from dejima-provider-lib.sh, so a caller loading only the latter -
+    # providers.bats does - validated against an empty set and rejected
+    # every guard there is.
+    grep -q '^DEJIMA_KNOWN_GUARDS=' "$REPO_ROOT/bin/dejima-provider-lib.sh"
+    run grep -n '^DEJIMA_KNOWN_GUARDS=' "$REPO_ROOT/bin/dejima-harness-lib.sh"
+    [ "$status" -ne 0 ] || { echo "the set is defined in two places"; return 1; }
+}
+
+@test "core gates guard work on the declaration, not on the plugin defining a hook" {
+    # A harness that forgot the declaration but implemented the hook would
+    # get a guard the launch message has just said it does not have.
+    grep -q 'harness_has_guard notify' "$REPO_ROOT/bin/dejima"
+}
+
+@test "install.sh installs no harness's hooks itself" {
+    # Where dejima's guards go, and in what shape, is the selected harness's
+    # business. install.sh knowing ~/.claude/settings.json's schema is the
+    # coupling; it delegates to the harness's plugin now.
+    run grep -nE '\.claude/(settings\.json|hooks)' "$REPO_ROOT/install.sh"
+    [ "$status" -ne 0 ] || { echo "install.sh still writes a harness's config:"; echo "$output"; return 1; }
+    grep -q 'harness_register_guards' "$REPO_ROOT/install.sh"
+    grep -q '^harness_register_guards()' "$REPO_ROOT/bin/harness/claude-code.sh"
+}
+
+@test "the plugin loader does not depend on its caller's SCRIPT_DIR" {
+    # bin/dejima sets it to bin/ and install.sh sets it to the repo root. A
+    # library that resolves paths from one of those only works for one of
+    # its callers.
+    sed -n '/^harness_plugin_load()/,/^}/p' "$REPO_ROOT/bin/dejima-harness-lib.sh" \
+        | grep -q 'BASH_SOURCE'
+    run sed -n '/^harness_plugin_load()/,/^}/p' "$REPO_ROOT/bin/dejima-harness-lib.sh"
+    [[ "$output" != *'$SCRIPT_DIR'* ]] || { echo "$output"; return 1; }
+}
+
+@test "a finding's label lives in trusted code, never in a manifest" {
+    # `ack` recovers an id by stripping a fixed label from the front of a
+    # finding. Those labels are a parser's grammar, and provider directories
+    # are not in tcb_snapshot's list - in the self-mount case a session can
+    # edit a manifest without raising a CRITICAL. So the label stays a
+    # literal in code the audit does watch, which is why host-exec surfaces
+    # are a plugin hook rather than a manifest key.
+    grep -q 'HOST-CODE-EXECUTION' "$REPO_ROOT/bin/harness/claude-code.sh"
+    run grep -rn 'HOST-CODE-EXECUTION' "$REPO_ROOT/harnesses"
+    [ "$status" -ne 0 ] || { echo "a finding label is manifest data:"; echo "$output"; return 1; }
+    run grep -rniE 'host_exec_paths|host_exec_label' "$REPO_ROOT/harnesses" "$REPO_ROOT/bin/dejima-provider-lib.sh"
+    [ "$status" -ne 0 ] || { echo "host-exec surfaces leaked into the manifest schema:"; echo "$output"; return 1; }
+}
+
+@test "the audit resolves host-execution paths from the harness, not a hardcoded one" {
+    run grep -n 'STATUSLINE_PATH' "$REPO_ROOT/bin/dejima-audit"
+    [ "$status" -ne 0 ] || { echo "$output"; return 1; }
+    grep -q 'host_exec_load' "$REPO_ROOT/bin/dejima-audit"
+    grep -q '^harness_host_exec_surfaces()' "$REPO_ROOT/bin/harness/claude-code.sh"
+}