Devbox Hardener
Bounds what a compromised development VM can reach on your network, and tells you what that bound actually is.
A devbox is a long-lived VM that runs dejima sessions. Unlike a CI runner it holds uncommitted work, so it cannot be rolled back on a schedule, and rollback is the whole containment story of the sister project, dejima-ci-hardener. What replaces it is the hypervisor firewall, plus knowing precisely what sits inside it.
It does not prevent root on the devbox. Nothing here tries to. It makes root reach less, and makes that reach something you can check rather than assume. What notices if it happens is dejima-guard, which is optional and lives on another machine.
| proxmox/README.md | Start here. The node side: create the VM, apply the firewall. Everything in the web console |
| docs/building-a-devbox.md | The guest side: install, bootstrap, verify. The same on any hypervisor |
| docs/threat-model.md | What each layer grants, what was rejected and why, and what is not bounded |
| Updating | Already have a devbox? Check its version and walk the upgrade path |
Do the node side first, through the end of its §3.
Before you use it
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 has the detail, and is where to report a vulnerability rather than in a public issue.
One person's work per devbox. A condition of the tool rather than an observation about how devboxes get used: the blast radius here is stated in terms of what is on the box, so a machine holding several people's work is out of spec and the reasoning does not hold for it. If you need separation between people, use a devbox each.
Container escape is deliberately deprioritised. An escape gains root, which is more than the container has, and root on this VM is already a row in the blast radius table, bounded by the same hypervisor firewall however it was reached. So patching is what is left worth doing, and separating escape from root inside the guest is not.
Use
sudo ./bin/dejima-devbox-bootstrap # once, in the guest
cd ~/dejima && ./install.sh # dejima's own step, as your own account
sudo dejima-devbox-audit # all three
sudo dejima-devbox-audit reach # containment only, what the timer runs
sudo dejima-devbox-audit radius # what root here would get
sudo dejima-devbox-audit drift # what changed since you last said it was clean
./bin/dejima-devbox-upgrade status # what this box is missing, from a checkout
Bootstrap installs what dejima needs from the host - Docker with the compose
v2 plugin, python3, jq - and clones dejima to ~/dejima. It stops there.
Installing dejima is dejima's step: install.sh writes into your ~/.claude,
~/.local/bin and ~/.config/dejima, and what it does there is that project's
to define and to change. A copy of that contract kept here would go stale
without anything noticing.
The harness is a step before that one, and neither this tool nor
install.sh arranges it: whichever harness dejima will run for you, it is
installed on this box as the account that will run dejima, and signed in, before
the first session. The credential proxy runs on this side and reads that
account's copy, and the container's mounts are that account's directories. docs/building-a-devbox.md
§5
has what each half breaks if it is missing.
The audit needs root. The config is 0640 root, and the radius report stats
files in the home of the account that runs dejima, which it reads from the
config rather than from $HOME - under sudo and under the timer that is
root's, whose home on a devbox holds nothing.
Updating an existing devbox
Check what it is at, from a checkout rather than the installed copy: an installed copy knows only the versions that existed when it was installed, and says so rather than reporting "up to date" about one it has never heard of.
git -C ~/dejima-devbox-hardener pull
cd ~/dejima-devbox-hardener
./bin/dejima-devbox-upgrade status
That names every version this box is missing. Do them in order: each note assumes the one before it, and some install what a later one uses.
| upgrades/ | One note per version: the console work, the guest work, and what it costs |
| docs/upgrading.md | The process, and why the two halves are counted separately |
Nothing here converges on its own, which is why any of this exists. A devbox is
pinned to the day it was built, half its controls were typed into a console no
git pull reaches, and a snapshot restore can move it backwards without anyone
deciding to. dejima needs none of it: short-lived containers, tip of main
only, and a pull closes any gap.
How it fails
The firewall is on the hypervisor, so the guest cannot verify it by reading it. That is what makes it worth having and what makes checking it awkward. The audit tests behaviour instead, by probing the default gateway: already in the routing table so it reveals nothing, guaranteed alive where an arbitrary address is not, and inside the ranges the firewall drops.
A timeout has two causes, and telling them apart is the whole design:
PASS |
Nothing answered and the network demonstrably works |
FAIL |
Something answered. Rules missing, or an exception above them |
INDETERMINATE |
Nothing answered, including things that should. Not a pass. |
A check that reports success when it does not know is worse than no check, because it gets believed.
The IPv6 half has the same three states for different reasons. It asks the
segment who is there over IPv6 and counts the answers, and there a silence is
unambiguous: the node dropped the family, or the guest never had it, and both
are the outcome wanted. So silence is a plain PASS with no positive control.
INDETERMINATE is reserved for the probe not running at all, which is still not
a pass. It is a separate check because the node's other rules name IPv4 ranges
and match none of that traffic - see
the LAN boundary is two boundaries.
A rule that REJECTs rather than DROPs reports FAIL even though it is
working. bash's /dev/tcp cannot tell a rejection from a live host
refusing. That is the safe direction to be wrong in, and it is why the template
specifies DROP.
Nothing is stored about what this VM must not reach. Such a list would be a map of valuable targets, kept on the machine this project assumes will be owned. The audit derives its targets from the routing table and its controls from your forge, so there is nothing to type and nothing to leak.
Nothing here is watching
Every check on this page runs on the box it is checking, when you run it. Root
on that box can lie to all of them, and the baseline drift compares against
lives on the machine whose drift it is describing. The audit bounds an intruder
who did not think to look. It has nothing to say about one who did, and nothing
at all to say at three in the morning.
dejima-guard is the other half, and it is a separate machine on purpose. It reaches in over SSH on a cadence, keeps the baseline on its own disk where root here cannot reach it, and alerts from there. The inversion is the point: a devbox that has been taken cannot silence a watcher it cannot talk to, and a collection that stops arriving is itself the alarm.
It needs nothing new from this box beyond v5: a restricted key that can run two read-only commands and nothing else, and one inbound firewall rule if the guard does not already sit on your admin subnet.
Tests
bats test/
test/devbox.bats asserts against a real devbox and skips everywhere else. It
is gated on DEJIMA_DEVBOX=1 rather than on detecting a devbox, because a
detector that guesses wrong skips silently, and a suite that passes by asserting
nothing is worse than one that fails.
Further reading
| docs/ | Index of everything below |
| docs/threat-model.md | What each layer grants, why escape is deprioritised, what was rejected |
| docs/reference.md | The three verdicts, the three addressing modes, the one config key |
| SECURITY.md | What this claims, and where to report a vulnerability |
| dejima-ci-hardener | The sister project, for hosts that can be rolled back |
| dejima-guard | Watches a devbox from a machine that is not the devbox, and alerts on what changed |
License
GPL-3.0-or-later. See LICENSE.
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, at dejima's ATTRIBUTION.md.