Add tools/ci/strict-evidence.py, which records only measured host/source
facts: tested SHA (git-verified when a checkout exists), source archive
digest, kernel, bwrap/rustc/cargo, uid, cgroup2, runner kind and outcome. The
lane now writes this document on every exit path, refuses a root-only PASS,
and covers the strict @knowledge-inert and forged-marker JSON contracts in
addition to the existing 7/7 adversarial matrix and positive CLI checks.
Rework security-linux.yml to run on trusted master pushes or an explicit
source_sha, bind the checkout to the tested SHA, create a git archive for the
digest, avoid PR execution and upload the evidence artifact. This is step 4 of
the CI/security RFC (B04/B09/B10); the runner must still be an ephemeral VM.
Strict mode now re-executes the CLI inside a Bubblewrap supervisor with
user/mount/PID/net/UTS/IPC namespaces, no-new-privs, all capabilities dropped,
a read-only /project, private /proc and /dev, rlimits, output capture and
wall-clock process-group cleanup. The child verifies PID 1, NoNewPrivs, empty
CapEff, an unshared uid map and loopback-only networking before project code.
- PlatformCapabilities::detect runs a real bwrap probe; no probe, no strict
- strict L1 supports pure Marqdo + Math; Foreign/ABI/worker/kernel/grants/trace
are rejected with exit 2 before sandboxing
- forged MARQDO_STRICT_ACTIVE is rejected; missing bwrap fails closed
- add adversarial helper, 7-test escape/resource matrix, seeded security lane
script and self-hosted security workflow
- executed G-C lane as uid 1000 in a disposable Ubuntu 26.04 VM: 7/7 PASS
plus positive CLI and refusal matrix; evidence in docs/evidence
- ADR-0021, spec/security.md, STATUS/COMPATIBILITY/README/CHANGELOG updated;
ADR-0019 remains for non-Linux/fail-closed rules