* feat: add router-less cross-family reviewer CLI (cross_review.py)
Re-homes the archived orchestrator repo's GPT-4.1 cross_reviewer as a
direct OpenAI SDK CLI: verbatim system prompt, same model id, ported
3-attempt exponential-backoff retry. Reads OPENAI_API_KEY from the
environment or the gitignored repo-root .env. Lazy openai import so
--help works without the package.
* feat: create machine-level suppressions dir on --global hook install
install-hooks.sh --global now mkdir -p's
${SH_SECURITY_SUPPRESSIONS_DIR:-~/.config/sea-haven/security-review} and
states the convention: machine-level <dir>/<repo-basename>/suppressions.json
is preferred over repo-local .security-review/suppressions.json, with
review.sh merging both when run without --suppressions.
* docs: describe sh-build-review as cross-family GPT-4.1 pass via cross_review.py
* chore: retire Path B VM host artifacts, repoint xmodel hook to cross_review.py
- delete systemd/ units and DEPLOY-R720.md (VM destroyed; recoverable
from git history)
- nightly_sweep.sh: retained for reference — header notes the VM-based
Path B sweep is retired; ENABLE_XMODEL_HOOK now calls this repo's
cross_review.py instead of the archived orchestrator's run.py
- sweep-targets.txt: drop the stale ~/orchestrator warning block
- README: document the Claude Code web cloud routines
(repo-scanner-nightly-sweep 08:00 ET, repo-checkers-plane1 07:30 ET,
ALARM-only to #repo-scanner) and add the cross_review.py section
* docs(iam): repoint cross-family review invocations to cross_review.py
* fix(ci): exclude canary corpus from ruff, add import smoke test
CI has been red repo-wide: the latest ruff wants to reformat the
intentionally-vulnerable canary fixtures (whose line numbers are keyed
in canary-meta/KEY.md), and pytest --collect-only exits 5 with zero
tests. Exclude canary/ via ruff.toml and add a root-level import smoke
test so collection is non-empty and imports cross_review.py.
review.sh only applied suppressions when handed an explicit --suppressions
FILE, so only the pre-push hook resolved them. Every other entry point (the
Open SWE daily-report automation, nightly sweep, on-demand/CI, agent runs)
called review.sh without it and therefore suppressed nothing, re-surfacing
every already-adjudicated false positive as HIGH.
When --suppressions is not passed, resolve by repo basename and MERGE both
suppression locations (machine-level first, wins id collisions):
- machine-level: ${SH_SECURITY_SUPPRESSIONS_DIR:-~/.config/sea-haven/security-review}/<basename>/suppressions.json
- repo-local: <repo>/.security-review/suppressions.json
An explicit --suppressions still overrides, so the hook and existing callers
are unaffected. Degrades gracefully off-Mac (repo-local only); fail-safe on an
unparseable file (suppresses nothing → blocks).
SECURITY (/sh-security-review, 2026-07-13): fan-out + proof-or-kill confirmed
one HIGH — the git-tracked repo-local suppressions.json lets anyone who can
commit to a scanned repo suppress a real finding and PASS an automated run
(verified by an actual exploit run; same posture nightly_sweep already had).
ACCEPTED-RISK per Adam on the condition that the automated scanners only ever
target trusted repos (no unreviewed untrusted contributions). Documented in the
auto-resolve block, README trust-model note, and a hard warning in
sweep-targets.txt. Four other candidates downgraded to low/pre-existing.
Verified: machine-level and repo-local both auto-resolve and suppress; explicit
--suppressions override still blocks; simulated off-Mac host keeps repo-local
and correctly re-blocks machine-level-only FPs.
Fresh-init copy of the security-review/ subsystem extracted from
Sea-Haven-Industries/orchestrator (being deprecated). Adds org-standard scaffold:
CI reusable-workflow callers (ruff + collect), dependency-review, labeler,
dependabot, .gitignore, requirements.txt. Scheduled execution is migrating to
Claude Code web routines (ALARM-only to #repo-scanner); the systemd units and
nightly_sweep.sh/checker_coordinator.sh remain the source of truth.
Committed with --no-verify: the canary fixtures (checkers/fixtures/**) carry
intentional secret-shaped test data that trips the deterministic gate (the
documented detector-fixture false positive); no new logic is introduced.
