fix(dictionary): repair the stat-list markup, drop the nested main landmarks, and finish the header sweep

[BigSimmo]

Summary

Follow-up to #2114, which merged while these two commits were still being written. They were pushed to that branch after the squash landed, so they never reached main; this PR replays them on the current base. No content is new since they were written — the branch was restarted from main and the two commits cherry-picked unchanged.

• definition-list (serious axe violation) — the /dictionary/sources stat strip put each tile's hint in a <p> beside the <dt>/<dd>, and a <div> inside a <dl> may contain only <dt>/<dd>. The hint moves inside the <dd>. /dictionary/sources is now axe-clean.
• Duplicate main landmarks on three routes — the topics list, the topic-detail terms column and the term page each rendered a second <main> inside the one InformationPageShell already provides, tripping landmark-no-duplicate-main, landmark-main-is-top-level and landmark-unique. Each becomes a div; no id, scroll anchor or nav target moves. This predates feat(dictionary): retitle the search route, strip page search bars, and rebuild the sources page #2114 — the repo's axe assertions only fail on critical/serious, so it never went red.
• Zero-result search — dropped its card and stood bare on the page background, reading as a failed render rather than an answer, and its advice named a filter even when none was applied. It now wears the same card as the result list, and the no-filter case offers Browse all terms.
• Zero-result topics — a list narrowed to nothing by a ?kind= or ?q= deep link showed "0 collections" above an empty rule and no explanation. It gets an empty state, and Clear now drops both narrowing inputs rather than only kind: clearing just kind left a ?q= visitor stuck on an empty page.
• Browse header — Browse was the one Dictionary route whose header feat(dictionary): retitle the search route, strip page search bars, and rebuild the sources page #2114 did not touch, so it still carried the descriptive line the other four lost ("Scan the same source-linked result system by letter or abbreviation."), leaving the mode inconsistent with itself. The sort control also drops ml-auto below sm: on a phone the toolbar wraps and a pushed-right sort landed alone on a second line, hanging off the right edge instead of reading as part of the control group.
• One queued ledger request (docs/outstanding-issues-inbox/91536913….json) recording the two axe findings that are not this mode's: region (3 nodes — the composer label, the privacy-warning span and the privacy link, all outside any landmark) and landmark-no-duplicate-banner/landmark-unique (the global #search universal header renders a second banner). Both reproduce on Dictionary routes this branch never touched, both live in shared chrome across all thirteen modes, and both are recorded rather than fixed here — a shared-chrome change under a Dictionary PR is the wrong blast radius.

Two headers are deliberately kept: the topic-detail description states that collection's scope (content, not decoration), and the Dictionary mode-home subtitle comes from the shared mode-home template every mode uses.

Verification

Re-run against this base after the replay, not carried over from the pre-merge branch.

• npm run lint — clean, --max-warnings 0
• npm run typecheck — clean
• npm run test — Test Files 673 passed (673) · Tests 7276 passed | 4 skipped (7280)
• npm run verify:ui when UI, routing, styling, browser behavior, reduced-motion, or forced-colors behavior changed — run as the affected Chromium journeys against an isolated production server: tests/ui-dictionary.spec.ts → 6 passed (27.3s), covering the 320/390/1440 overflow sweep and the axe pass under dark + forced colors.
• Additional evidence: a direct @axe-core/playwright sweep of the six Dictionary routes at 1440px confirms /dictionary/sources is clean and the three landmark rules no longer fire on /dictionary/topics, /dictionary/topics/[slug] or /dictionary/[slug]. Screenshots taken at 390px and 1440px for the changed states.
• Verification not run: npm run verify:pr-local — its heavy scope repeats the lint, typecheck and unit results pasted above, and the browser evidence for this diff is the journey set named.
• Verification not run: npm run eval:retrieval:quality, npm run eval:rag, npm run eval:quality, npm run check:production-readiness, npm run verify:release — all provider-backed, and no retrieval, ranking, selection, chunking, scoring, answer-generation or environment behaviour changed.

Risk and rollout

• Risk: presentation and markup only. Three main elements become divs, one <p> moves inside its <dd>, two empty states gain copy and a card, and one header line is removed. No route, handler, data path or contract changes. The landmark change makes getByRole("main") less ambiguous rather than more.
• Rollback: git revert either commit independently; they touch overlapping files but not overlapping hunks.
• Provider or production effects: None.

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

No clinical content changes. The sources page keeps its "Source linked is not specialist clinical approval" callout, its approval-pending wording, and its per-source region and accessed dates verbatim; the stat tiles remain derived from dictionary-data.ts rather than authored. The two new empty states assert nothing about clinical content — they explain why a list is empty and offer a way back. No SaMD-relevant decision-support behaviour changed.

RAG impact: no retrieval behaviour change — no file under src/lib/rag/**, clinical-search, retrieval-selection, ranking-config, answer-ranking, the eval harness or the golden fixture is touched; the diff is four presentation files plus one queued ledger request.

Notes

• Why this is a second PR: feat(dictionary): retitle the search route, strip page search bars, and rebuild the sources page #2114 squash-merged at 12:18 UTC while the branch was at its second commit. These two commits were pushed at 12:43 and 13:32 UTC — after the merge — onto a branch whose PR was already closed, so they could never have merged from there. The branch was restarted from main and the commits cherry-picked; the pre-restart tip is kept locally as backup/dictionary-orphans-1f69abe.
• The queued inbox request is merge-safe and applies to the canonical ledger through npm run issues:reconcile from a serialized branch after this lands.

---

Generated by Claude Code

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[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 ↗︎.

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your current included review allowance is based on your included PR review attempts over the past 7 days.

Next review available in:34 minutes

Limit details: You’ve used the included review currently available. Your 101 included PR review attempts over the past 7 days set your current allowance at 1 review per hour.

Your organization has reached its usage spending cap. Adjust your spending cap in the billing tab.

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits within each organization.

For paid Pro and Pro+ reviews, CodeRabbit uses a developer's included PR review attempts over the past 7 days to set the current hourly allowance. At typical activity levels, the full plan allowance applies. Higher sustained activity can lower the allowance until earlier attempts leave the 7-day window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 7969a9c7-03a2-43f1-a4c8-99b023e1093d

📥 Commits

Reviewing files that changed from the base of the PR and between baec777 and 839334a.

📒 Files selected for processing (6)

• docs/branch-review-records/196a129d152e41b7d7af1e82b2a335d905520acf60450bcfb3925cd089df6c32.record.md
• docs/branch-review-records/1c99cd1f4f7fc28a1bae4d20b72bac77c80665af8487cac6441b92e183bf5922.record.md
• docs/outstanding-issues-inbox/91536913-6cd9-4c61-a109-4d9aa3e76b9d.json
• src/components/dictionary/dictionary-catalogue-pages.tsx
• src/components/dictionary/dictionary-sources-page.tsx
• src/components/dictionary/dictionary-term-page.tsx

---

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