Sanitise Therapy copy for the record, and stop the two apt tiers racing

[BigSimmo]

Summary

Two unrelated fixes, each the smallest that closes the cause.

1. Therapy copy reached the clipboard unsanitised. The Therapy corpus carries 1,956 U+2192 arrows across body, patientExplanation, deliverySteps and briefVersion, plus en/em dashes and curly quotes. The record systems this text is pasted into render those as replacement glyphs or drop them silently.

plainClinicalText already solves exactly this on the DSM note path, so this reuses it rather than rewriting the data. Rewriting was the wrong instrument: the 205 therapy records are review-tracked by a content digest (therapyReviewedContentSha256, documented as "changing any non-review field invalidates sign-off"), so editing their prose is a clinical governance act, not a formatting pass. Copy is the boundary where the text stops being a web page and becomes note content, so it is sanitised there — in copyText, the single choke point all four Therapy Compass screens already share.

plainClinicalText moves to src/lib/plain-clinical-text.ts with no data imports, because dsm-note.ts imports the DSM corpus and the therapy bundle must not pull that in. dsm-note.ts re-exports it, so every existing caller and tests/dsm-note.test.ts are untouched.

2. The cloud setup script's browsers tier failed whenever it ran alongside the python tier. Both drive apt — browsers via playwright install --with-deps, python via apt_install for tesseract — and the per-tier lock deliberately does not serialise across tiers. Session mode runs them in one detached child while instructing the model to confirm with bash scripts/setup-claude-cloud.sh browsers python, so the documented confirmation command is precisely what collides. The 2026-09-06 log carries apt's own explanation:

E: dpkg was interrupted, you must manually run 'dpkg --configure -a' to correct the problem.

Playwright surfaced that only as Installation process exited with code: 100, and the browsers tier was reported failed even though nothing about it was broken. A shared apt lock now serialises the two, waiting rather than failing, since apt is busy for a bounded time.

Verification

• npm run verify:cheap — fully green. Test Files 1266 passed | 1 skipped (1267), Tests 18015 passed | 2 expected fail | 3 skipped (18020), exit 0. Lint, typecheck and all static gates clean.
• npm run typecheck and npx eslint on the changed sources — clean.
• Lock behaviour proven, not assumed. The with_apt_lock function was extracted verbatim from the script under test and driven by two concurrent callers. Output shows strict serialisation with no interleaving:

  [t] waiting for another run's apt step to finish
  START B 1788772212234100441
  END   B 1788772214238218531
  START A 1788772217244875627
  END   A 1788772219249096878
  

• bash -n scripts/setup-claude-cloud.sh clean. shellcheck is not installed in this environment, so it was not run.
• Two new cases in tests/therapy-compass-clipboard.dom.test.tsx pin that therapy copy spells out arrows, thresholds, dashes and quotes, and leaves plain clinical text unchanged.
• PR policy failure on 53ebfdb reproduced and fixed locally against evaluatePullRequestPolicy itself: the first body reworded three governance items and the check compares them by exact string equality. Old wording reproduces ok: false with that one error; this body returns ok: true with zero errors and zero warnings.
• UI verification not run: npm run plan:browser failed closed to the full suite ("No spec references anything rendered by src/components/therapy-compass/use-clipboard.ts"), and CI runs the complete Chromium suite on this change.
• npm run verify:pr-local not run: verify:cheap above is the broader offline gate and came back green on this exact commit.

Risk and rollout

• Risk: Low. Change 1 alters only what reaches the clipboard on Therapy Compass, using a transformation already trusted on the DSM note path; no data, schema or review record is touched. Change 2 is provisioning-script only and cannot affect the application.
• Rollback: Revert this single commit; both changes are independent and additive.
• Provider or production effects: None / no provider call, deploy, or migration is involved.
• RAG impact: no retrieval behaviour change — no file under src/lib/rag/, no retrieval RPC, ranking config, eval harness or golden fixture is touched; the diff is a clipboard sanitiser and a shell provisioning lock.

Clinical Governance Preflight

• Source-backed claims still require linked source verification before clinical use
• No patient-identifiable document workflow was introduced or expanded without explicit governance approval
• Supabase target remains Clinical KB Database (sjrfecxgysukkwxsowpy)
• Service-role keys and private document access remain server-only
• Demo/synthetic content remains clearly separated from real clinical sources
• Source metadata, review status, and outdated/unknown-source behavior remain conservative
• Deployment classification/TGA SaMD impact was checked when clinical decision-support behavior changed

Notes

• On governance item 6: no therapy record content, reviewStatus, or review digest is modified by this PR, so no existing or future clinician sign-off is invalidated. That is the reason the fix sits at the copy boundary rather than in the data.
• All 205 therapy records are currently needs_review with no reviewedBy or reviewedContentSha256 set, so rewriting the source would not have invalidated anything today. It is still the wrong place to fix this: sanitising at the copy boundary keeps the reviewed text and the pasted text separable once sign-off does begin (outstanding item #SBKXZ7).
• On governance item 7: no clinical decision-support behaviour changes, so the deployment/TGA classification is unchanged by this diff.
• The sanitiser also folds semicolons to commas and em dashes to hyphens on this path, which is the existing documented behaviour of plainClinicalText, not a new decision.

🤖 Generated with Claude Code

https://claude.ai/code/session_01TX21C93LAT68nNvCYHWTsG

---

Note

Low Risk
Clipboard-only formatting on Therapy copy reuses an existing DSM sanitizer; cloud changes affect provisioning shell only and cannot touch production app behavior.

Overview
Therapy Compass copy now runs clipboard text through plainClinicalText in the shared copyText helper, so arrows, thresholds, dashes, and curly quotes are spelled out for EMR paste targets without editing therapy records. The sanitizer is extracted to src/lib/plain-clinical-text.ts (no DSM data imports) and re-exported from dsm-note.ts so existing DSM callers stay unchanged.

Claude cloud provisioning adds with_apt_lock so apt_install and playwright install --with-deps cannot hit dpkg concurrently when session background provisioning overlaps the documented browsers python confirmation run. The lock waits (configurable timeout, stale lock reclaim); on timeout the tier fails rather than running apt unlocked.

New tests cover apt lock serialisation (script-sourced harness) and therapy clipboard sanitisation; diff-integrity test count for claude-cloud-profile.test.ts is bumped to 26.

^{Reviewed by Cursor Bugbot for commit c543085. Configure here.}

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Team

Run ID: 6f9940a0-285d-481c-a97b-a6153360f2a1

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[supabase]

This pull request has been ignored for the connected project sjrfecxgysukkwxsowpy because there are no changes detected in supabase directory. You can change this behaviour in Project Integrations Settings ↗︎.

---

Preview Branches by Supabase.
Learn more about Supabase Branching ↗︎.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-07T09:53:11.931379Z	53ebfdb	Manual request

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_a8c7b4da-8336-49dc-992a-d63d3da36f0b)

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 53ebfdb7da

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".

[BigSimmo]

@codex resolve actionable Codex review findings for this pull request and current head using the repository instructions. This is the pull request's single automatic repair pass: do not perform a fresh review, create new standalone findings, or request another review. Work only the existing unresolved Codex threads on the current head. The only repository destination is BigSimmo/Database, and the only branch destination is the pull request head branch claude/eager-maxwell-z0h0z0 at starting commit 53ebfdb; never publish fixes to a detached or synthetic work branch and never create a stacked pull request. Use the authenticated GitHub connector to commit each approved fix to BigSimmo/Database:claude/eager-maxwell-z0h0z0, then verify that the pull request head contains the published commit before reporting success. Always fix P0 and P1 findings. For P2 and lower findings, fix only clear, scoped, low-risk issues; otherwise disposition them with a concise reason. For a fixed thread, reply with as the first line and as the second line. For a no-code disposition, use followed by . These result markers authorize the workflow to close that exact thread only after it verifies a fixed commit is the pull request head; a local-only commit is not a fix. If publication or verification fails, do not use either result marker, do not claim success, and leave the thread open with the blocker. If human input or new authorization is required, do the same. Finish only after every actionable thread is fixed or dispositioned and closed, or explicitly left open for a human decision. Do not update the branch from main, address unrelated reviews, broaden scope, or create more than one scoped fix commit. Do not use external APIs, paid services, credentials, dependency changes, or broad refactors unless explicitly authorized. Add targeted tests where behavior changes and run the narrowest relevant validation.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Bravo.

Reviewed commit: 53ebfdb7da

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".

[github-actions]

CI triage

CI failed on this PR. Automated classification of the 2 failed job(s):

• Unit coverage — needs investigation: inspect the failing step and uploaded diagnostics; rerun only after classifying the cause.
• PR required — needs investigation: inspect the failing step and uploaded diagnostics; rerun only after classifying the cause.

Compared with main CI run #16283 (cancelled).

Classification is evidence routing, not permission to ignore a failure. Exact quarantined Playwright identities remain governed by the flake ledger.

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_b3f6ea88-7303-4b69-bde7-01a8fa4a57fa)