Derive the forms PDF password badge from committed bytes and gate it - #2531
Merged
Conversation
`data/forms-pdf-manifest.json` carries a `passwordProtected` flag per WA Mental Health Act statutory form, and the form detail page renders a clinician-facing badge from it. The manifest was entirely hand-maintained — no generator, no validator, no gate — and the existing test compared the flag only against itself plus one hardcoded literal for Form 12A, so a wrong flag on any of the other 50 forms would have passed silently. `scripts/build-forms-pdf-manifest.mjs` now derives sha256, bytes and passwordProtected from the committed bytes, offline, with a `--check` mode registered as `check:forms-pdf-manifest` in `verify:cheap:internal` and CI. The flag is derived by ATTEMPTING to open each PDF with an empty user password, not by looking for an `/Encrypt` marker. A PDF encrypted with an owner password but no user password carries `/Encrypt` and still opens freely, so the marker answers a different question than the badge asks. Every committed form happens to agree under both rules, which is exactly why the weaker rule would have survived review: a synthetic owner-password-only fixture is now the discriminating case, and it is the only test that fails if the deriver is ever simplified to a grep. Every error path — unreadable file, truncated body, malformed xref, unparseable encryption dictionary, unsupported revision — yields `passwordProtected: true` and a hard non-zero exit, never a silent `false`. `false` is the assertion that a clinician can open the form, so under-warning is the unsafe direction. `--check` fails on a manifest/bytes disagreement rather than auto-correcting, because the manifest also carries the sha256 provenance record. The generator never synthesises an `officialPdfUrl`, never reorders entries, hard-errors on a PDF with no entry, and now pins the publisher host rather than accepting any https URL. The manifest data itself is unchanged and regenerates byte-identically. The badge reads "Password required to open". It previously read "Password protected"; an interim wording claimed the password comes from the publisher, which nothing here establishes — the bytes prove only that opening requires a non-empty user password, not who holds it or whether a clinician can obtain it. On a statutory form that difference is the difference between a warning and a false errand at the bedside. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
Append-only inbox request recording that the forms PDF manifest now derives its password flag from committed bytes and gates it. Separate commit so it stays independently revertible from the product change. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
Contributor
|
Important
This repository does not receive automatic reviews because it has fewer than 10 stars. ⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Team Run ID: 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. Comment |
|
This pull request has been ignored for the connected project Preview Branches by Supabase. |
…ent title Verified in Chromium at a 320px viewport, A/B against the same page with only this string changed. In the PDF row the badge and the title share a two-column grid, and the badge is not allowed to shrink: "Password protected" (main) -> title/subtitle track 70px "Password required to open" -> title/subtitle track 35px "Password required" -> title/subtitle track 78px At 35px the document title rendered as "Rec…" and its publisher line as "Offic…". That is a real regression against main, so the longer wording is not worth its cost: it squeezed the name of the statutory form in order to describe the lock on it. "Password required" keeps everything the bytes establish — opening the file needs a non-empty user password — and still makes no claim about who holds it, which was the reason the earlier "…from the publisher" wording was rejected. It is also one character shorter than the string on main, so the row is slightly better off than before this branch. No layout or tone class changed; the fix is entirely in the wording. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
4 tasks
BigSimmo
marked this pull request as ready for review
September 2, 2026 05:57
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. |
Contributor
Bugbot couldn't run - usage limit reachedBugbot 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_54a1303f-ba21-444f-9acd-1a2cfdfdebdc) |
![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: 111a5ee089
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".
Codex review, verified. `buildManifest` rebuilt each asset by enumerating six keys, so any other field an asset carried was dropped on regeneration. The generator computes exactly three fields — sha256, bytes, passwordProtected — and everything else is provenance it cannot recompute. This script is offline by contract, so a dropped `officialPdfUrl` could never be restored from the PDFs themselves. The failure mode is quiet and then destructive: `--check` reports drift, and the regeneration it instructs the operator to run erases the field for good. The code already disagreed with itself, which is the clearest evidence the finding is real. The unreadable-file path a few lines above spreads `...asset` and only overrides the derived fields; the success path did not. They now match. Not hypothetical: an `editingRestricted` fact derived from the PDFs' /P permission bits is already a queued follow-up against this same manifest, so the first field added would have hit this. Pinned by a test that adds two extra properties to an asset and asserts they survive while the three derived fields stay authoritative. Negative control: restoring the enumerated-keys version fails it with `expected undefined to be '2026-09-02'`. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
Contributor
Bugbot couldn't run - usage limit reachedBugbot 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_dd107e8d-b785-4ec7-bce1-964182954a63) |
Resolves the generated docs/scripts-index.md conflict: main's build-and-assets section plus this branch's build-forms-pdf-manifest.mjs entry, with the counts sentence regenerated by scripts/update-docs-inventory.mjs rather than hand-edited. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
Same generated-counts conflict as the previous sync: docs/scripts-index.md's "(N files) / (N entries)" sentence is rewritten by scripts/update-docs-inventory.mjs, so both sides are regenerated output. Resolved by regenerating (285 files, 288 npm scripts) rather than picking a side, with this branch's build-forms-pdf-manifest.mjs entry preserved. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
Contributor
Bugbot couldn't run - usage limit reachedBugbot 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_084c88bb-0e98-44e2-af93-040ad20f9c7e) |
BigSimmo
enabled auto-merge (squash)
September 2, 2026 09:15
BigSimmo
pushed a commit
that referenced
this pull request
Sep 2, 2026
…Y090R5 sibling) The generated counts sentence at the top of docs/scripts-index.md conflicts on every concurrent PR — measured five times on PR #2531 in about three hours, at five different main heads, each time as the sole conflicting line. Same root cause as #Y090R5 for data/outstanding-issues-snapshot.json: a single-line generated artefact every PR must regenerate. PR #2530 fixed the snapshot half by moving regeneration into the serialised reconcile step; this half is untouched. Records the two traps found while resolving it: taking main's whole file silently drops the branch's own new script entry, and a bundled verify:cheap:internal union merge desynchronises the gate counts that check:gate-manifest reads. Append-only inbox request; no canonical ledger edit. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
Two conflicts, both from concurrent additions rather than disagreement: package.json — main added check:diff-integrity to verify:cheap:internal while this branch added check:forms-pdf-manifest. Neither removed anything, so the resolution is the union: both gates in the chain, each in its own position. That in turn desynchronised the gate counts check:gate-manifest compares against, so CLAUDE.md (35 -> 36 static) and .claude/skills/gates/SKILL.md (38 -> 39 total) were re-derived from the merged chain, not hand-guessed. docs/scripts-index.md — the generated counts sentence, for the fifth time on this branch. Regenerated with scripts/update-docs-inventory.mjs (286 files, 290 npm scripts) rather than picking a side, with this branch's build-forms-pdf-manifest.mjs entry preserved. Queued as a ledger request on claude/issues-followups: it is the same single-line-generated-artefact collision that #Y090R5 records for the issues snapshot. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
…aude/form-12a-warning
Contributor
Bugbot couldn't run - usage limit reachedBugbot 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_95e9ac53-38a1-4cd7-bfa5-75ca13c018ff) |
Sixth sync of the generated docs/scripts-index.md counts sentence on this branch. Regenerated with scripts/update-docs-inventory.mjs (287 files, 292 npm scripts) rather than picking a side; this branch's build-forms-pdf-manifest.mjs entry preserved. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
10 tasks
BigSimmo
pushed a commit
that referenced
this pull request
Sep 2, 2026
…form PDFs
A concurrent review comment on this PR asserted that PyMuPDF does not enforce
/P permission bits and that these files therefore "open normally and yield
their text layer". Measured against the committed bytes with PyMuPDF 1.28.0,
the library the worker actually uses, that is false for 50 of the 51 files.
needs_pass=1, is_encrypted=True, page_count=0
doc.authenticate("") -> 0 (the empty user password is rejected)
load_page(0) -> ValueError('document closed or encrypted')
form-12a.pdf is the sole exception: no /Encrypt, opens, 3274 characters of
first-page text — which is exactly the file PR #2531 corrected.
The distinction the comment missed is that /P alone would not block opening;
these files also carry a non-empty /U, and that does. The practical
consequence is that extraction fails at the open call, before
should_ocr_page() is ever reached, so the OCR fallback cannot rescue them.
The JavaScript fallback was not measured and is not claimed either way.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
BigSimmo
added a commit
that referenced
this pull request
Sep 4, 2026
…r review (#2544) * issues: queue seven verified follow-ups from PRs #2538, #2536 and #2531 Seven immutable inbox requests, no canonical ledger edit. Every claim was checked against the code before it was written: three against origin/main, four against the PR heads the follow-up belongs to (#2538 ef7c55a, #2536 e24e0ee, #2531 60f5e8e), since none of the three has merged yet. - P2 issue: differential-records.ts asserts validation_status "locally_reviewed" from a literal over a snapshot that says "Pending review" — the sibling of the fix already in medication-records.ts. - P2 issue: registry-records.ts and differential-records.ts both return the frozen source_status column verbatim, and derive it with a substring test that also matches "not checked"/"unchecked". - P3 issue: a stored source_status of "outdated" can never be cleared, since nothing writes that column back on any of the three record tables. - P3 rec: patient-alert rows with action "info" reach neither unassessed tier, leaving a green all-clear for nystatin, levetiracetam and lorazepam. - P3 issue: isProfileEmpty treats a recorded hepatic "none" as no information, disagreeing with the engine, which treats it as an answer. - P3 task: the considerations panel's two not-assessed sentences format their input lists differently. - P3 rec: the forms PDF manifest records only passwordProtected, though the committed bytes also say modification, text extraction and assembly are blocked while printing and form-filling are permitted. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL * issues: queue the recurring docs/scripts-index.md counts collision (#Y090R5 sibling) The generated counts sentence at the top of docs/scripts-index.md conflicts on every concurrent PR — measured five times on PR #2531 in about three hours, at five different main heads, each time as the sole conflicting line. Same root cause as #Y090R5 for data/outstanding-issues-snapshot.json: a single-line generated artefact every PR must regenerate. PR #2530 fixed the snapshot half by moving regeneration into the serialised reconcile step; this half is untouched. Records the two traps found while resolving it: taking main's whole file silently drops the branch's own new script entry, and a bundled verify:cheap:internal union merge desynchronises the gate counts that check:gate-manifest reads. Append-only inbox request; no canonical ledger edit. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL * issues: cancel the OCR-routing claim and replace it with a corrected record Codex review finding on this PR, verified against the code it cited rather than taken on trust. The finding is correct. Request e3133c1b asserted that disabled text extraction is "exactly the condition that sends the worker down the OCR fallback path". It is not. should_ocr_page() in worker/python/extract_pdf_assets.py (line 186) decides on extracted text length and image coverage ratio only, and never reads the /P permission bits. The extractor has no needs_pass or authenticate handling at all, so a user-password PDF fails before any OCR decision is reached. A wrong causal mechanism in a durable record is worse than no record: the next person plans ingestion work from it. Cancelled rather than edited, because inbox requests are immutable by design and the cancel action exists for exactly this — the audit trail keeps the wrong claim, its refutation, and the correction. The permission-bits finding itself stands unchanged and is restated in the replacement request, which makes no ingestion claim and says plainly that what happens to these files on ingestion is untested. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL * issues: record the reproducible mobile /documents/search LCP breach Two confirmed measurements on different heads of PR #2536 — +466ms (+20.4%) and +471ms (+20.6%) against a +20%/+100ms tolerance, five milliseconds apart. That is reproducible, not noise, and the passing re-run between them was the outlier. Ruled out as PR #2536's doing: its diff is six medication files, and documents/search/page.tsx imports one symbol (Metadata from next). Nothing outside medication-named files calls /api/medications. Recorded as a hypothesis, not a finding: lighthouse-budget.json was last refreshed 2026-08-27 and main has taken heavy change since, so the PR carrying the newest main absorbs the blame. Same disease as #QSHHGK for the bundle budget. Main itself was never measured, and the record says so. Needs an owner decision — find the regression, or refresh the baseline on a schedule rather than reactively to clear a red PR. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL * issues: cancel the Lighthouse regression claim and replace it with the measured variance A third graded run refutes the "reproducible regression" reading in request d1b3491f. CI run 33684986161 (head 957a038) measured mobile /documents/search at 2323ms against the 2282ms baseline — +41ms — and reported "Every graded route is within tolerance of the committed baseline." The cell has now produced two breaches near 2750ms and at least two passes near 2320ms on the same pinned Chromium and the same route. That is a bimodal measurement whose two modes straddle the +20%/+100ms tolerance, not a page that became half a second slower. The gate's 2-of-3 sampling already tolerates noise within a run; this split is between runs, which that design does not cover. Cancelled rather than edited, because inbox requests are immutable and the cancel action exists for this. The replacement states all four outcomes and asks for the variance to be characterised — repeat the dispatch-only baseline-refresh job against main and compare the spread — rather than for a regression hunt or a reactive baseline refresh. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL * issues: record the measured encryption state of the committed WA MHA form PDFs A concurrent review comment on this PR asserted that PyMuPDF does not enforce /P permission bits and that these files therefore "open normally and yield their text layer". Measured against the committed bytes with PyMuPDF 1.28.0, the library the worker actually uses, that is false for 50 of the 51 files. needs_pass=1, is_encrypted=True, page_count=0 doc.authenticate("") -> 0 (the empty user password is rejected) load_page(0) -> ValueError('document closed or encrypted') form-12a.pdf is the sole exception: no /Encrypt, opens, 3274 characters of first-page text — which is exactly the file PR #2531 corrected. The distinction the comment missed is that /P alone would not block opening; these files also carry a non-empty /U, and that does. The practical consequence is that extraction fails at the open call, before should_ocr_page() is ever reached, so the OCR fallback cannot rescue them. The JavaScript fallback was not measured and is not claimed either way. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL * issues: record the intermittent caring-contacts-guidance strict-mode violation Observed once on PR #2536 (CI run 33690849576 attempt 1, Production UI shard 2) and cleared by a single re-run of the same commit: strict mode violation: getByTestId("caring-contacts-guidance") resolved to 2 elements The source renders that test id in exactly one place, and both the page and the shell interpolate it once, so the duplicate is not a second render site. The likely window is React relocating out-of-order streamed content from the page's next/dynamic shell, which the test's waitUntil:"load" does not wait past — but that is a hypothesis and the record says so; the retained trace should confirm it before anyone edits the test. Not PR #2536's: its diff is medication and documentation files only, and the same shard passed on its previous head. The record asks for the three unscoped locators to be scoped to the main landmark, and explicitly rules out quarantining — one reproduction is below the repository's three-on-the-same-SHA bar, and a locator fix is not a suppression. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL * docs(ledger): record Run PR sweep review for PR #2544 Merged origin/main into claude/issues-followups (clean, no conflicts) and verified the narrow gates for this append-only inbox PR; both review threads were already resolved from a prior pass. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0137etKe6X2p9HkXsXoYY8TA --------- Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: BigSimmo <87357024+BigSimmo@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
data/forms-pdf-manifest.jsoncarries apasswordProtectedflag per WA Mental Health Act statutory form, and the form detail page renders a clinician-facing badge from it. The manifest was entirely hand-maintained: no generator, no validator, no gate. The existing test compared the flag only against itself, plus one hardcoded literal for Form 12A, so a wrong flag on any of the other 50 forms would have passed silently.scripts/build-forms-pdf-manifest.mjswith a--checkmode, following thebuild-therapies-index.mjs/check:mha-act-sectionsidiom, and registerscheck:forms-pdf-manifestinverify:cheap:internaland CI./Encryptmarker. A PDF encrypted with an owner password but no user password carries/Encryptand still opens freely, so the marker answers a different question than the badge asks. Every committed form happens to agree under both rules — which is exactly why the weaker rule would have survived review — so a synthetic owner-password-only fixture is now the discriminating case, and it is the only test that fails if the deriver is ever simplified to a grep.passwordProtected: trueand a hard non-zero exit, never a silentfalse. Under-warning is the unsafe direction:falseis the assertion that a clinician can open the form.--checkfails on a manifest/bytes disagreement rather than auto-correcting, because the manifest also carries the sha256 provenance record. The generator never synthesises anofficialPdfUrl, never reorders entries, hard-errors on a PDF with no entry, and pins the publisher host rather than accepting any https URL.The manifest data itself is unchanged and regenerates byte-identically; this PR adds the derivation and the gate that keep it honest.
Two review rounds changed the badge copy before merge, in opposite directions. An interim wording read "Opens with a password from the publisher"; a clinical governance review blocked it, because the bytes establish only that opening requires a non-empty user password — not who holds it, whether the publisher issues it, or whether a clinician can obtain it. On a statutory form that is the difference between a warning and a false errand at the bedside. The replacement, "Password required to open", was then measured in a browser and found to squeeze the document title (below); "Password required" says everything the bytes support and fits.
The
#9P4XAEclosure request rides as its own separately revertible commit.Verification
npm run test(full unit suite) —Test Files 947 passed (947)/Tests 12058 passed | 1 skipped (12059)node scripts/build-forms-pdf-manifest.mjs --check—Forms PDF manifest is current (51 PDFs, 50 require a user password).tests/forms.test.ts+tests/source-catalogue-providers.test.ts—Test Files 2 passed (2)/Tests 27 passed (27)(re-run after the badge change; still 27 passed)npm run typecheck—recorded a pass for "typecheck:internal" (6001 input files)npm run check:gate-manifest— "all 38 verify:cheap gates are enforced in CI (static-pr + mapped jobs), and the 35 static gates are documented consistently"npm run check:ledger-write-discipline— "Ledger write discipline passed for 45a3dca..HEAD"npx eslinton the changed source files — exit 0npm run format(whole tree) — no files changedRegeneration without
--checkproduced a byte-identical manifest, which is the evidence that the committed data was already correct under the stricter rule.Two negative controls were run. Flipping Form 12A's flag to
truemade the gate exit 1 and the new derived-from-bytes assertion fail with12A: expected false to be true. Replacing the deriver with a naive/Encryptgrep made exactly the two relevant tests fail —expected { passwordProtected: true, …(1) } to deeply equal { passwordProtected: false, …(1) }on the owner-password fixture, plus the truncated-input case in the fail-closed test. Both were reverted and the suite re-verified green.npm run test:focusedrefuses by design here ("Focused test selection is unsafe: test or configuration paths changed") and instructs the full suite, so the full suite is what ran.Two failures seen on an earlier run of this branch (
tests/clinical-hazard-controls.test.ts,tests/rag-plan-package-parity.test.ts) were a shallow-clone artefact — the container held 98 commits and both tests name commits absent from it. Aftergit fetch --deepen=2000the suite is green as quoted above, with no code change.Verification not run:
npm run check:production-readinesscannot pass in this repository.check:privacy-readiness:releaseblocks on six release-blocking legal/provider items (OpenAI ZDR, OpenAI and Railway DPAs, APP 8 cross-border basis, APP 1/APP 5 notice, PHI minimisation). The same command on unmodifiedmainat45a3dca, with no changes, produces identical output — this is repository state, not a consequence of this diff, and none of those items is touched here.Verification not run:
npm run verify:pr-localwas not run; its risk-routed selection adds no failure class this diff can reach beyond the full unit suite already run above.UI verified in a browser.
npm run ensure(server identity confirmed via/api/local-project-id), then Chromium at 320px and 360px on/forms/form-10aand/forms/form-12a. No horizontal page overflow at either width.The first attempt at this badge was measurably worse than
mainand was fixed before merge. In the PDF row the badge shares a two-column grid with the document title and is not allowed to shrink, so an A/B at 320px changing only that string gave:At 35px the statutory form's title rendered as "Rec…" and its publisher line as "Offic…". The shipped wording leaves the row slightly better off than
main.UI verification not run: the full Chromium journey gate
npm run verify:uiwas not run because the change is one string with no layout, tone, routing or tap-target change, and the affected row was inspected directly.Risk and rollout
main, so a revert removes the generator and the check without any data change to repair.Clinical Governance Preflight
Clinical KB Database(sjrfecxgysukkwxsowpy)Notes
scripts/pr-policy.mjswarns that operational-risk changes are bundled with UI changes here. That is unavoidable rather than incidental:check:gate-manifestenforces a one-way invariant that CI must never run less of theverify:cheapstatic set than the local chain does, so a new local gate cannot be added without registering it in.github/workflows/ci.ymlin the same commit. TheCLAUDE.md(34 → 35) and.claude/skills/gates/SKILL.md(37 → 38) edits are the documented gate counts the same check compares against, anddocs/scripts-index.mdis generated inventory. No policy text was altered — only numerals and one curated script entry.This branch is two commits behind
main(569d8565,a1aa449a).git merge-tree --write-tree origin/mainreports a clean merge.569d8565restructured the instruction docs, but the two gate-count sentencescheck:gate-manifestreads still live inCLAUDE.mdand.claude/skills/gates/SKILL.mdon currentmain, and the newdocs/agents/verification-gates.mdcarries no count — so this branch's numerals still land in the right place after merge. Behind, not conflicted, and deliberately not re-synced to avoid restarting CI for nothing.🤖 Generated with Claude Code
https://claude.ai/code/session_0183EiexuZe6uKqoACXGuANL
Note
Low Risk
Manifest JSON is unchanged when regenerated; the change adds validation and conservative derivation plus a small UI label tweak, with no auth, data migration, or network surface.
Overview
Adds an offline
build-forms-pdf-manifest.mjspipeline sodata/forms-pdf-manifest.jsonis no longer trusted by hand forsha256,bytes, andpasswordProtected.passwordProtectedis decided by trying to open each committed PDF with an empty user password (pdf.js), not by grepping/Encrypt; corrupt or unclassifiable files fail closed to protected plus a hard exit.check:forms-pdf-manifest(--check) is wired intoverify:cheap:internaland CI static-heavy checks so drift againstpublic/forms-pdf/blocks merge.tests/forms.test.tsnow pins every flag to the bytes, fail-closed error paths, manifest field preservation on regen, and an owner-password-only synthetic PDF so a naive/Encryptshortcut would fail. The form detail badge copy changes from "Password protected" to "Password required". Docs/gate counts bump to 36 static gates; outstanding issue #9P4XAE is closed in the inbox ledger.Reviewed by Cursor Bugbot for commit 3fcd6cf. Configure here.