Harden RAG retrieval and answer reliability by BigSimmo · Pull Request #1289 · BigSimmo/Database

BigSimmo · 2026-07-27T23:12:47Z

Summary

Harden RAG evidence binding, clinical-value verification, comparison fallback, extractive formatting, citation preservation, and fail-closed behavior across current clinical fixtures.
Resolve #019 admission/discharge source loss and #029 review-fallback stubs, with a zero-tolerance quality gate for recurrence.
Improve eval diagnostics, irrelevant-at-10 evidence, source-governance reporting, and path-specific coverage floors.
Add the auditable BMJ third-party attestation policy, schema/API support, migration, and pending-human-review manifest without treating BMJ content as clinically verified.
Reconcile the RAG backlog and evidence documents; lithium is closed, while ADHD/metabolic evidence debt, browser-matrix evidence, and human label review remain open.

RAG impact: behaviour change — canary pair output/rag-retrieval-post-final.json -> output/rag-retrieval-post-exact-head.json (36/36; document/content recall 1.0; zero per-case document/content RR regressions)

The protected retrieval change uses the expanded chunk query only for the measured clozapine blood-count action shape. Discharge and other red-section document lookups retain raw intent terms. Ranking scores, comparator order, clamps, and semantic reranking are unchanged.

Verification

Exact-head typechecking passed within npm run verify:pr-local.

Risk and rollout

Risk: High clinical/reliability surface. Main risks are incorrect evidence selection, misplaced numeric values, lost citations, overconfident extraction, and schema/API version skew. New fixtures and fail-closed gates cover these paths.
Rollback: Revert the application changes. The Supabase migration is currently unapplied, so no database rollback is required. If applied later, leave the additive schema in place and use a separately reviewed forward migration rather than dropping data.
Provider or production effects: Authorized live evals and read-only hosted RPC profiling were performed. No hosted migration, source attestation, or production-data write was performed. Merging to main will follow the repository's normal application/worker auto-deploy path.

Migration 20260727010000_bmj_third_party_source_attestation.sql is a separate, deliberate operator step and must not be applied automatically. Until applied, legacy owner-scoped review behavior remains available and v2-only public/attestation requests fail closed with 503 source_review_v2_unavailable.

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

Supersedes draft Harden RAG retrieval and answer reliability #1288 with a clean one-commit history so the deliberately fake token-shaped test fixture is absent from every PR commit.
Current reconciled head: 22b7b751247f1904b11bd53145ac18fbb0330b90.
BMJ sources remain clinical_validation_status=unverified; qualified human review, hosted apply, attestation, and warning-rate remeasurement remain #022.
Firefox/WebKit evidence and human irrelevant-at-10 disposition remain #023.
ADHD corpus/table accessibility and metabolic schedule evidence remain the open parts of #018.

coderabbitai · 2026-07-27T23:12:53Z

Warning

Review limit reached

@BigSimmo, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 2 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

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 for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: d9b5237d-5dc2-4425-ad6d-3765fd3e203f

📥 Commits

Reviewing files that changed from the base of the PR and between b9340a3 and 22b7b75.

📒 Files selected for processing (76)

docs/branch-review-ledger.md
docs/evidence/rag-reliability-evidence-2026-07-27.md
docs/evidence/rag-top-local-review-manifest-2026-07-26.json
docs/outstanding-issues.md
docs/rag-behaviour/refuted-approaches.md
docs/rag-hybrid-findings-and-todo.md
docs/source-governance-refresh-worklist-2026-07-22.md
scripts/audit-source-governance.ts
scripts/build-clinical-review-queue.ts
scripts/eval-answer-quality.ts
scripts/eval-quality.ts
scripts/eval-retrieval.ts
scripts/eval-utils.ts
scripts/profile-retrieval-rpcs.ts
src/app/api/documents/[id]/reviews/route.ts
src/app/api/documents/bulk/route.ts
src/lib/answer-ranking.ts
src/lib/answer-verification.ts
src/lib/clinical-review-queue.ts
src/lib/clinical-search.ts
src/lib/clinical-value-binding.ts
src/lib/eval-document-matching.ts
src/lib/medication-entities.ts
src/lib/rag/rag-answer-text.ts
src/lib/rag/rag-cache.ts
src/lib/rag/rag-candidate-sources.ts
src/lib/rag/rag-claim-support.ts
src/lib/rag/rag-comparison.ts
src/lib/rag/rag-eval-cases.ts
src/lib/rag/rag-extractive-answer.ts
src/lib/rag/rag-extractive-artifacts.ts
src/lib/rag/rag-extractive-first.ts
src/lib/rag/rag-quote-verification.ts
src/lib/rag/rag-routing.ts
src/lib/rag/rag-source-segmentation.ts
src/lib/rag/rag.ts
src/lib/retrieval-selection.ts
src/lib/smart-rag-api.ts
src/lib/source-review.ts
src/lib/supabase/database.types.ts
supabase/drift-manifest.json
supabase/migrations/20260727010000_bmj_third_party_source_attestation.sql
supabase/schema.sql
tests/answer-ranking.test.ts
tests/answer-responsiveness-gate.test.ts
tests/answer-verification.test.ts
tests/clinical-review-queue.test.ts
tests/clinical-search.test.ts
tests/document-mutation-routes.test.ts
tests/eval-answer-quality.test.ts
tests/eval-document-matching.test.ts
tests/eval-quality.test.ts
tests/eval-retrieval.test.ts
tests/eval-utils.test.ts
tests/extractive-answer-formatting.test.ts
tests/medication-entities.test.ts
tests/profile-retrieval-rpcs.test.ts
tests/rag-answer-fallback.test.ts
tests/rag-answer-text.test.ts
tests/rag-claim-support.test.ts
tests/rag-document-lookup-chunk-query.test.ts
tests/rag-document-lookup-escalation-rescue.test.ts
tests/rag-eval-cases.test.ts
tests/rag-extractive-first.test.ts
tests/rag-extractive-procedural-artifact.test.ts
tests/rag-fast-path-ordering.test.ts
tests/rag-offline-answer.test.ts
tests/rag-routing.test.ts
tests/rag-trust.test.ts
tests/retrieval-selection.test.ts
tests/smart-rag-api.test.ts
tests/source-authority-tooling.test.ts
tests/source-review-policy.test.ts
tests/source-review-route.test.ts
tests/source-review-schema.test.ts
vitest.config.mts

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch codex/rag-reliability-final

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

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

supabase · 2026-07-27T23:12:59Z

Updates to Preview Branch (codex/rag-reliability-final) ↗︎

Deployments	Status	Updated
Database	✅	Tue, 28 Jul 2026 00:36:40 UTC
Services	✅	Tue, 28 Jul 2026 00:36:40 UTC
APIs	✅	Tue, 28 Jul 2026 00:36:40 UTC

Tasks are run on every commit but only new migration files are pushed.
Close and reopen this PR if you want to apply changes from existing seed or migration files.

Tasks	Status	Updated
Configurations	✅	Tue, 28 Jul 2026 00:36:41 UTC
Migrations	✅	Tue, 28 Jul 2026 00:36:43 UTC
Seeding	✅	Tue, 28 Jul 2026 00:36:45 UTC
Edge Functions	✅	Tue, 28 Jul 2026 00:36:46 UTC

View logs for this Workflow Run ↗︎.
Learn more about Supabase for Git ↗︎.

…-final

chatgpt-codex-connector · 2026-07-28T00:12:33Z

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.

…-final # Conflicts: # docs/outstanding-issues.md

BigSimmo · 2026-07-28T00:26:45Z

@cursoragent run bugbot review

cursor · 2026-07-28T00:26:51Z

Bugbot review of PR #1289 — APPROVE (exact head d49af8ac; ledger commit now at 94deee52)

Findings

No high-confidence P0–P2 defects.

Severity	Location	Finding	Disposition
—	—	No P0/P1/P2	No defect

Application src/ / supabase/ / tests/ / scripts/ are unchanged vs the prior APPROVE tip a1ca6a01. Tip delta was docs evidence + a clean merge from main. No unresolved cursor[bot] Bugbot threads on the PR.

Residual risk

Unapplied BMJ attestation migration (#022) — public/attestation reviews correctly hard-fail with 503 source_review_v2_unavailable until a qualified hosted apply.

Checks

Focused high-risk Vitest: 358 + 1502 passed
merge-tree CLEAN vs origin/main
Not run: live canaries / provider-backed gates (same app tree as prior recorded canary)

Ledger updated and pushed: #1289

BigSimmo · 2026-07-28T00:27:10Z

@claude Review this PR

github-actions · 2026-07-28T00:27:18Z

Hi @BigSimmo — thanks for the @claude mention.

Open this in Claude Code on the web to continue here: https://claude.ai/code

Context: BigSimmo/Database · #1289

_{Automated backlink from the Claude backlink workflow. Claude Code can pick up the mention and push a scoped fix to the branch.}

claude · 2026-07-28T00:27:33Z

Claude encountered an error after 2s —— View job

I'll analyze this and get back to you.

Copilot

Pull request overview

This PR hardens high-risk RAG retrieval/answer reliability and source-governance surfaces by tightening evidence binding, improving deterministic/extractive fallbacks, strengthening eval diagnostics + quality gates, and introducing BMJ third‑party attestation support that remains explicitly unverified pending qualified human action.

Changes:

Strengthens retrieval selection + document lookup chunk selection to preserve intent terms and prevent wrong-subject/table-evidence promotion (notably around clozapine blood-count action shapes and lithium monitoring rows).
Tightens answer/eval reliability contracts (required expected-file citations, explicit handling of source-backed review stubs, richer retrieval diagnostics persisted into artifacts).
Adds BMJ third-party attestation policy plumbing across SQL schema/types/tests and admin-only review route behavior that fails closed when the v2 RPC/migration is unavailable.

Reviewed changes

Copilot reviewed 75 out of 76 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
vitest.config.mts	Adds aggregate coverage floors for core RAG/reliability modules.
tests/source-review-schema.test.ts	Verifies BMJ attestation migration/schema/types invariants and fail-closed constraints.
tests/source-authority-tooling.test.ts	Pins audit/backfill tooling to include source_review_events + BMJ attestation handling.
tests/smart-rag-api.test.ts	Ensures extractive comparison planning is described as clinical synthesis with deterministic assembly.
tests/retrieval-selection.test.ts	Adds regression cases preventing foreign-medication/foreign-table promotion and preserving subject binding.
tests/rag-offline-answer.test.ts	Tightens assertion to require the correct clozapine withhold threshold signal and rejects review-stub routing.
tests/rag-fast-path-ordering.test.ts	Guards stable ordering for lithium monitoring rows over higher-scored unrelated tables/foreign rows.
tests/rag-eval-cases.test.ts	Adds/strengthens eval case contracts for authoritative citations and review-stub scoring behavior.
tests/rag-document-lookup-escalation-rescue.test.ts	Pins the smallest-sufficient rescued evidence instead of incentivizing citation padding.
tests/rag-document-lookup-chunk-query.test.ts	Covers the new chunk-query selection behavior (expanded variant only for measured clozapine action shape).
tests/rag-answer-text.test.ts	Extends answer-quality heuristics around time-unit phrases (“6 hours”).
tests/medication-entities.test.ts	Adds medication entity safety/catalog coverage for foreign-medication detection and binding.
tests/eval-utils.test.ts	Adds required-claim + expected-citation binding validation coverage.
tests/eval-retrieval.test.ts	Adds assertions for persisted per-row relevance/signal matches in golden retrieval artifacts.
tests/eval-quality.test.ts	Adds CLI arg validation, safe dumping constraints, and gating behavior for unsupported-only runs.
tests/eval-document-matching.test.ts	Tightens overlap rules with explicit approvals and document-identity dedupe for matching.
tests/document-mutation-routes.test.ts	Prevents bulk mutation routes from writing governance status or editing reviewed identity/provenance.
tests/clinical-search.test.ts	Adds foreign-threshold label handling and subject-binding regression coverage.
tests/clinical-review-queue.test.ts	Extends review queue with authority classification and deterministic top-local evidence manifest.
tests/answer-responsiveness-gate.test.ts	Ensures timeout fallback text doesn’t echo user typos.
tests/answer-ranking.test.ts	Ensures numeric ranges stay within a single bold span.
supabase/schema.sql	Adds source_review_events decision, new columns, and record_source_review_v2 with service-role-only grants.
supabase/drift-manifest.json	Updates schema snapshot/hash including new columns/RPC signature.
src/lib/supabase/database.types.ts	Exposes new source_review_events columns + record_source_review_v2 args.
src/lib/source-review.ts	Implements BMJ attestation policy constants, eligibility checks, and event+metadata completeness verification.
src/lib/smart-rag-api.ts	Adjusts intent for multi-document synthesis to reflect extractive-but-synthesis behavior.
src/lib/retrieval-selection.ts	Expands FBC matching and centralizes monitoring subject matching via clinical-search helpers.
src/lib/rag/rag-source-segmentation.ts	Adds bounded line reflow helpers and atomic NMHS clozapine red-range segment recovery.
src/lib/rag/rag-quote-verification.ts	Hardens citation enrichment to only add deterministic support when model-selected citations are directly supported and conflict-free.
src/lib/rag/rag-extractive-artifacts.ts	Detects procedural comparator artifacts (ECT flow edge/arrow glyphs) without blanket comparator rejection.
src/lib/rag/rag-eval-cases.ts	Adds requiredDirectClaim contract + scoring changes to reject source-backed review stubs as “covered”.
src/lib/rag/rag-comparison.ts	Adds safe comparison fallback selection that respects bounded fail-closed families.
src/lib/rag/rag-candidate-sources.ts	Narrows expanded chunk-query usage to the measured clozapine blood-count action shape; otherwise preserves raw intent query.
src/lib/rag/rag-cache.ts	Bumps cache dependency version to invalidate stale behavior.
src/lib/rag/rag-answer-text.ts	Prevents “dose 6 hours” false positives while still flagging missing units.
src/lib/eval-document-matching.ts	Adds document-identity-based dedupe + maximum matching to prevent one physical doc satisfying multiple expected slots.
src/lib/clinical-value-binding.ts	Adds explicit entity↔value binding helper for clause-scoped safety checks.
src/lib/clinical-review-queue.ts	Adds review class, best-rank tracking, and top-local evidence manifest builder.
src/app/api/documents/bulk/route.ts	Disallows bulk governance status writes and fails closed on identity/provenance edits for reviewed sources.
src/app/api/documents/[id]/reviews/route.ts	Adds BMJ attestation inputs, v2 RPC routing + fail-closed 503 when migration is unapplied, and public-doc cache invalidation.
scripts/eval-utils.ts	Implements required-direct-claim semantic contract validation bound to expected-file citations.
scripts/eval-retrieval.ts	Persists per-result relevance grade + matched declared signals into golden artifacts for diagnostics.
scripts/build-clinical-review-queue.ts	Adds `--top-local-manifest` + `--source-artifact` support and emits deterministic manifest/report.
scripts/audit-source-governance.ts	Adds service-role-only source_review_events audit to compute operational unattested debt without emitting event evidence.
docs/source-governance-refresh-worklist-2026-07-22.md	Updates operator worklist with implemented BMJ policy status + manifest evidence boundaries.
docs/rag-hybrid-findings-and-todo.md	Updates backlog items with landed bounded OCR repair + latency profiling evidence.
docs/rag-behaviour/refuted-approaches.md	Documents additional refuted approaches and updated irrelevant@10 diagnostics posture.
docs/evidence/rag-top-local-review-manifest-2026-07-26.json	Adds auditable top-local evidence manifest (pending qualified human review, no attestation applied).
docs/evidence/rag-reliability-evidence-2026-07-27.md	Records local/live evidence paths and outcome metrics for the reliability work.
docs/branch-review-ledger.md	Appends a review ledger entry for PR #1289 / `codex/rag-reliability-final`.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Append an append-only ledger row for tip d49af8a after a Bugbot pass found no P0–P2 on the unchanged application tree. Co-authored-by: BigSimmo <BigSimmo@users.noreply.github.com>

…-final

BigSimmo · 2026-07-28T00:35:47Z

Pull request overview

This PR hardens high-risk RAG retrieval/answer reliability and source-governance surfaces by tightening evidence binding, improving deterministic/extractive fallbacks, strengthening eval diagnostics + quality gates, and introducing BMJ third‑party attestation support that remains explicitly unverified pending qualified human action.

Changes:

Strengthens retrieval selection + document lookup chunk selection to preserve intent terms and prevent wrong-subject/table-evidence promotion (notably around clozapine blood-count action shapes and lithium monitoring rows).

Tightens answer/eval reliability contracts (required expected-file citations, explicit handling of source-backed review stubs, richer retrieval diagnostics persisted into artifacts).

Adds BMJ third-party attestation policy plumbing across SQL schema/types/tests and admin-only review route behavior that fails closed when the v2 RPC/migration is unavailable.

Reviewed changes

Copilot reviewed 75 out of 76 changed files in this pull request and generated no comments.

Show a summary per file

| File | Description |

| ---- | ----------- |

| vitest.config.mts | Adds aggregate coverage floors for core RAG/reliability modules. |

| tests/source-review-schema.test.ts | Verifies BMJ attestation migration/schema/types invariants and fail-closed constraints. |

| tests/source-authority-tooling.test.ts | Pins audit/backfill tooling to include source_review_events + BMJ attestation handling. |

| tests/smart-rag-api.test.ts | Ensures extractive comparison planning is described as clinical synthesis with deterministic assembly. |

| tests/retrieval-selection.test.ts | Adds regression cases preventing foreign-medication/foreign-table promotion and preserving subject binding. |

| tests/rag-offline-answer.test.ts | Tightens assertion to require the correct clozapine withhold threshold signal and rejects review-stub routing. |

| tests/rag-fast-path-ordering.test.ts | Guards stable ordering for lithium monitoring rows over higher-scored unrelated tables/foreign rows. |

| tests/rag-eval-cases.test.ts | Adds/strengthens eval case contracts for authoritative citations and review-stub scoring behavior. |

| tests/rag-document-lookup-escalation-rescue.test.ts | Pins the smallest-sufficient rescued evidence instead of incentivizing citation padding. |

| tests/rag-document-lookup-chunk-query.test.ts | Covers the new chunk-query selection behavior (expanded variant only for measured clozapine action shape). |

| tests/rag-answer-text.test.ts | Extends answer-quality heuristics around time-unit phrases (“6 hours”). |

| tests/medication-entities.test.ts | Adds medication entity safety/catalog coverage for foreign-medication detection and binding. |

| tests/eval-utils.test.ts | Adds required-claim + expected-citation binding validation coverage. |

| tests/eval-retrieval.test.ts | Adds assertions for persisted per-row relevance/signal matches in golden retrieval artifacts. |

| tests/eval-quality.test.ts | Adds CLI arg validation, safe dumping constraints, and gating behavior for unsupported-only runs. |

| tests/eval-document-matching.test.ts | Tightens overlap rules with explicit approvals and document-identity dedupe for matching. |

| tests/document-mutation-routes.test.ts | Prevents bulk mutation routes from writing governance status or editing reviewed identity/provenance. |

| tests/clinical-search.test.ts | Adds foreign-threshold label handling and subject-binding regression coverage. |

| tests/clinical-review-queue.test.ts | Extends review queue with authority classification and deterministic top-local evidence manifest. |

| tests/answer-responsiveness-gate.test.ts | Ensures timeout fallback text doesn’t echo user typos. |

| tests/answer-ranking.test.ts | Ensures numeric ranges stay within a single bold span. |

| supabase/schema.sql | Adds source_review_events decision, new columns, and record_source_review_v2 with service-role-only grants. |

| supabase/drift-manifest.json | Updates schema snapshot/hash including new columns/RPC signature. |

| src/lib/supabase/database.types.ts | Exposes new source_review_events columns + record_source_review_v2 args. |

| src/lib/source-review.ts | Implements BMJ attestation policy constants, eligibility checks, and event+metadata completeness verification. |

| src/lib/smart-rag-api.ts | Adjusts intent for multi-document synthesis to reflect extractive-but-synthesis behavior. |

| src/lib/retrieval-selection.ts | Expands FBC matching and centralizes monitoring subject matching via clinical-search helpers. |

| src/lib/rag/rag-source-segmentation.ts | Adds bounded line reflow helpers and atomic NMHS clozapine red-range segment recovery. |

| src/lib/rag/rag-quote-verification.ts | Hardens citation enrichment to only add deterministic support when model-selected citations are directly supported and conflict-free. |

| src/lib/rag/rag-extractive-artifacts.ts | Detects procedural comparator artifacts (ECT flow edge/arrow glyphs) without blanket comparator rejection. |

| src/lib/rag/rag-eval-cases.ts | Adds requiredDirectClaim contract + scoring changes to reject source-backed review stubs as “covered”. |

| src/lib/rag/rag-comparison.ts | Adds safe comparison fallback selection that respects bounded fail-closed families. |

| src/lib/rag/rag-candidate-sources.ts | Narrows expanded chunk-query usage to the measured clozapine blood-count action shape; otherwise preserves raw intent query. |

| src/lib/rag/rag-cache.ts | Bumps cache dependency version to invalidate stale behavior. |

| src/lib/rag/rag-answer-text.ts | Prevents “dose 6 hours” false positives while still flagging missing units. |

| src/lib/eval-document-matching.ts | Adds document-identity-based dedupe + maximum matching to prevent one physical doc satisfying multiple expected slots. |

| src/lib/clinical-value-binding.ts | Adds explicit entity↔value binding helper for clause-scoped safety checks. |

| src/lib/clinical-review-queue.ts | Adds review class, best-rank tracking, and top-local evidence manifest builder. |

| src/app/api/documents/bulk/route.ts | Disallows bulk governance status writes and fails closed on identity/provenance edits for reviewed sources. |

| src/app/api/documents/[id]/reviews/route.ts | Adds BMJ attestation inputs, v2 RPC routing + fail-closed 503 when migration is unapplied, and public-doc cache invalidation. |

| scripts/eval-utils.ts | Implements required-direct-claim semantic contract validation bound to expected-file citations. |

| scripts/eval-retrieval.ts | Persists per-result relevance grade + matched declared signals into golden artifacts for diagnostics. |

| scripts/build-clinical-review-queue.ts | Adds --top-local-manifest + --source-artifact support and emits deterministic manifest/report. |

| scripts/audit-source-governance.ts | Adds service-role-only source_review_events audit to compute operational unattested debt without emitting event evidence. |

| docs/source-governance-refresh-worklist-2026-07-22.md | Updates operator worklist with implemented BMJ policy status + manifest evidence boundaries. |

| docs/rag-hybrid-findings-and-todo.md | Updates backlog items with landed bounded OCR repair + latency profiling evidence. |

| docs/rag-behaviour/refuted-approaches.md | Documents additional refuted approaches and updated irrelevant@10 diagnostics posture. |

| docs/evidence/rag-top-local-review-manifest-2026-07-26.json | Adds auditable top-local evidence manifest (pending qualified human review, no attestation applied). |

| docs/evidence/rag-reliability-evidence-2026-07-27.md | Records local/live evidence paths and outcome metrics for the reliability work. |

| docs/branch-review-ledger.md | Appends a review ledger entry for PR #1289 / codex/rag-reliability-final. |

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

@cursoragent Any issues identified from copilot review?

BigSimmo · 2026-07-28T00:36:47Z

@CodeRabbit please review

coderabbitai · 2026-07-28T00:36:57Z

@BigSimmo: I’ll review the changes, with particular attention to fail-closed behavior, evidence/citation binding, retrieval regressions, and migration/API compatibility.

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

Remove exact duplicate branch-review-ledger rows introduced when merging origin/main so Static PR checks stay green.

Union-merge of concurrent PR #1289 rows produced exact duplicates. Restore docs/branch-review-ledger.md from origin/main and re-append only this PR's disposition rows. Co-authored-by: BigSimmo <BigSimmo@users.noreply.github.com>

feat(rag): harden retrieval and answer reliability

eebe938

BigSimmo mentioned this pull request Jul 27, 2026

Harden RAG retrieval and answer reliability #1288

Closed

22 tasks

BigSimmo added 2 commits July 28, 2026 07:28

Merge remote-tracking branch 'origin/main' into codex/rag-reliability…

a1ca6a0

…-final

docs(rag): record exact-head release evidence

b7fd210

BigSimmo marked this pull request as ready for review July 28, 2026 00:12

Merge remote-tracking branch 'origin/main' into codex/rag-reliability…

d49af8a

…-final # Conflicts: # docs/outstanding-issues.md

BigSimmo requested a review from Copilot July 28, 2026 00:25

Copilot started reviewing on behalf of BigSimmo July 28, 2026 00:26 View session

Copilot AI reviewed Jul 28, 2026

View reviewed changes

cursoragent and others added 2 commits July 28, 2026 00:30

docs: record Bugbot exact-head review for PR #1289

94deee5

Append an append-only ledger row for tip d49af8a after a Bugbot pass found no P0–P2 on the unchanged application tree. Co-authored-by: BigSimmo <BigSimmo@users.noreply.github.com>

Merge remote-tracking branch 'origin/main' into codex/rag-reliability…

ab6ca03

…-final

docs(rag): record final main sync review

22b7b75

BigSimmo merged commit ec04f17 into main Jul 28, 2026
19 checks passed

BigSimmo deleted the codex/rag-reliability-final branch July 28, 2026 00:50

cursor Bot pushed a commit that referenced this pull request Jul 28, 2026

fix(ledger): dedupe union-merge duplicates after #1289 sync

091b433

Remove exact duplicate branch-review-ledger rows introduced when merging origin/main so Static PR checks stay green.

cursor Bot mentioned this pull request Jul 28, 2026

feat(ui): codebase review improvements and design system ratcheting #1286

Merged

This was referenced Jul 28, 2026

perf(ui): prefetch mode homes for seamless switching #1275

Merged

fix: fail-closed on numeric claim truncation overflow #1292

Merged

This was referenced Jul 28, 2026

test(search): validate focused live document search #1320

Merged

fix: high-confidence bugs from last-100-PR review #1374

Merged

Conversation

BigSimmo commented Jul 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Verification

Risk and rollout

Clinical Governance Preflight

Notes

Uh oh!

coderabbitai Bot commented Jul 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Review limit reached

Uh oh!

supabase Bot commented Jul 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

chatgpt-codex-connector Bot commented Jul 28, 2026

Uh oh!

BigSimmo commented Jul 28, 2026

Uh oh!

cursor Bot commented Jul 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Findings

Residual risk

Checks

Uh oh!

BigSimmo commented Jul 28, 2026

Uh oh!

github-actions Bot commented Jul 28, 2026

Uh oh!

claude Bot commented Jul 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

BigSimmo commented Jul 28, 2026

Pull request overview

Reviewed changes

Uh oh!

BigSimmo commented Jul 28, 2026

Uh oh!

coderabbitai Bot commented Jul 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

BigSimmo commented Jul 27, 2026 •

edited

Loading

coderabbitai Bot commented Jul 27, 2026 •

edited

Loading

supabase Bot commented Jul 27, 2026 •

edited

Loading

cursor Bot commented Jul 28, 2026 •

edited

Loading

claude Bot commented Jul 28, 2026 •

edited

Loading

coderabbitai Bot commented Jul 28, 2026 •

edited

Loading