Repository navigation
Security and compliance: bot auto-merge gate, worker scan window, TGA limbs (d) and (e) - #3742
Conversation
…XSQ) PR policy now fails a clinical-risk PR while auto-merge on it was armed by any account other than the repository owner. This enforces the 2026-10-20 decision that Gitar never arms clinical-risk PRs (gitar-bot armed PR 3507, which then merged on bot approvals alone). The owner may still arm it. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
|
Warning Review limit reachedYour organization has reached its usage spending cap. Adjust your spending cap in the billing tab. Next included review available in 17 minutes. View limit details
Comment |
|
This pull request has been ignored for the connected project Preview Branches by Supabase. |
![gitar-bot[bot]](https://avatars.githubusercontent.com/in/827041?s=60&v=4){kind=small}
…decision to the clinician Owner decisions 2026-10-21: - HIGH/CRITICAL findings in the worker image are advisory on main until 4 Nov 2026 so a red worker scan stops holding up deploys. App image findings and unassessed scans still fail; the window ends by itself. - Both answer prompts gain the TGA limb (d) line: no diagnosis or treatment decision for a specific patient. Legacy prompt-cache revision rolled. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
BigSimmo
changed the title
… records Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
…mb e) Medium-trust answers now show the exact sentences relied on, capped at 2. Authority-capped medium answers still show none. DEMO_ANSWER_CONFIDENCE=medium renders supported demo answers at medium trust for local screenshots with no provider call. Built for owner screenshot review; merges only after approval. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
BigSimmo
changed the title
…ng owner screenshot approval Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
…l of medium-trust quotes The local screenshot check showed no visible difference with demo content, so the switch served no one. Owner approved the change on 21 Oct. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
…-5zy0n0 # Conflicts: # tests/trivy-image-scan.test.ts
|
Canary pair: "before" baseline recorded. This is Eval Canary run 39630009061 on main
The "after" run follows the merge, because the canary only runs from main. It must show doc and content recall still at 1.00, no per-case rr regression, and no new failed answer cases. If any regress, this PR is reverted at once. Generated by Claude Code |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
BigSimmo has reached the 50-credit limit for trial accounts. To continue receiving code reviews, upgrade your plan.
|
Claude encountered an error after 2s —— View job I'll analyze this and get back to you. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 831ff94ac9
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
A release branch push or a manual dispatch from another ref now blocks on worker HIGH or CRITICAL findings, as the Codex review asked. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5
Code Review ✅ Approved🔴 High risk · Making worker-image HIGH/CRITICAL findings advisory on main until Nov 4 could let vulnerable images pass CI and reach production. Adds a PR policy gate that blocks bot-armed auto-merge on clinical-risk PRs, a time-limited advisory window for HIGH/CRITICAL worker image scan findings until 4 Nov 2026, prompt lines for TGA exemption limb (d), and up to 2 exact quote cards at medium trust for limb (e). No issues found. VerificationRules No rules evaluated Functional validation Not enabledAuto-approval Approved For more verification details, comment OptionsAuto-apply is off → Gitar will not commit updates to this branch. Comment with these commands to change the behavior for this request:
Was this helpful? React with 👍 / 👎 | Gitar |
|
PR sweep (22 Oct): I've turned off auto-merge here. This PR changes owner sign-off areas (PR policy, seals, clinical content or visible UI). The PR body summarises the owner's approvals, but there is no owner approval written in the owner's own words on the PR. Auto-merge stays off until the owner posts one. Sweep agents won't re-arm it. |
|
Lighthouse budget failed on b634c8b: desktop-root LCP +217 ms (+22.7%) and TBT +76 ms against the baseline. This PR doesn't touch the home page or any client bundle. The last commit only changed Generated by Claude Code |
|
Moved to draft by Grok Bot while it waits for the owner's own-words approval, because gitar-bot keeps re-arming auto-merge on disarmed PRs. Mark ready again once he posts approval here. |
There was a problem hiding this comment.
BigSimmo has reached the 50-credit limit for trial accounts. To continue receiving code reviews, upgrade your plan.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b634c8bec4
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
…exceptions Ran npm run governance:seal-hazard-controls, which reseals the reviewed digests for src/lib/answer-render-policy.ts and tests/answer-render-policy.test.ts and clears the two driftExceptions entries, per the owner's decision and the Codex P1 review thread.
|
I approve this PR, including the medium-trust quote cards, the TGA prompt line and the worker scan window to 4 Nov." |
|
"I approve resealing the clinical hazard controls on #3742 and removing the path-wide drift exception." |
|
Static PR checks failed on 833f34b at Generated by Claude Code |
|
Canary pair, "after" run: no regression. No revert. Eval Canary 39742639771 ran on main
Generated by Claude Code |
Requested by Josh · project thread
Summary
Four changes. The owner approved each one on 21 Oct 2026, first with "Go ahead with all recommendations please" and then the quote cards and reseal with "Go ahead with, 2 and 3 and 4 5 and 6". On 22 Oct the owner also approved them on this PR in his own words, including the reseal.
docs/clinical-hazard-controls.jsonwas resealed (833f34b) after the owner's approval. That records the reviewed digests forsrc/lib/answer-render-policy.tsand its test and removes the temporary drift exceptions.Areas touched: Delivery and assurance, Knowledge and records, Answers
RAG impact: behaviour change. The answer prompt text and the render caps changed, but retrieval and ranking did not. The owner approved one before and after eval canary pair, capped at US$2. The canary only runs from main, so the before run goes on main before merge and the after run follows the merge. A regression in the after run means an immediate revert of this PR.
Verification
node scripts/pr-policy.mjs --self-testpassed, including five new cases.npx vitest run tests/pr-policy*.test.tspassed 39 tests.tests/trivy-image-scan.test.ts,tests/container-ci-contract.test.tsandtests/ci-audit-contracts.test.tspassed 148 tests, including the advisory-window and main-only cases.tests/answer-render-policy.test.tspassed. The prompt and answer tests (rag-prompt-injection,rag-answer-composition-prompt,rag-answer-fallback) passed. The client payload, clinical safety, trust, dashboard helper, demo boundary and decision tests passed.check:clinical-hazard-controlspassed after the reseal.check:env-parityandcheck:repo-awareness-snapshotpassed. The pre-push static guard (lint and typecheck) passed.Risk and rollout
Clinical Governance Preflight
PsychSift Production(sjrfecxgysukkwxsowpy)The prompt line and the quote cards keep the amended TGA exemption limbs (d) and (e) from 1 Nov 2026. See docs/clinical-hazard-analysis.md, "Amended exemption from 1 November 2026".
🤖 Generated with Claude Code
https://claude.ai/code/session_01W7EHqBKdEqicWVBdYqv4W5