fix(review): bound impact-map query cache retention

[JSONbored]

Motivation

• Prevent unbounded storage growth from attacker-controlled, per-file impact-map query fingerprints by ensuring expired cache rows do not accumulate.
• Remove retaining large retrieved context blobs in the cache since computeImpactMap only needs metrics.paths for rendering.
• Make repo-scoped eviction efficient via an index so periodic cleanup operations can target stale rows without full-table scans.

Description

• Evict stale rows on cache reads and proactively delete rows older than the TTL before inserting a fresh result by adding a cutoff helper and DELETE ... WHERE fetched_at < ? in putCachedImpactMapQuery and deleting the expired single-row on stale getCachedImpactMapQuery misses in src/review/impact-map.ts.
• Stop storing the retrieved context body for impact-map entries and write an empty context payload while preserving metrics_json in putCachedImpactMapQuery so per-entry storage cost is minimized.
• Add a (project, repo, fetched_at) index in the raw migration and Drizzle schema so repo-scoped stale-row deletes are efficient (migrations/0132_impact_map_query_cache.sql, src/db/schema.ts).
• Extend the cachingStorageStub and add two regression unit tests that assert expired rows are evicted and cached rows keep only metrics (no context) in test/unit/impact-map.test.ts.
• Regenerate worker runtime types to satisfy the cf-typegen check (worker-configuration.d.ts).

Testing

• Ran the focused unit tests with npx vitest run test/unit/impact-map.test.ts and all impact-map tests passed (24 tests passed).
• Verified schema and type drift with npm run db:schema-drift:check, npm run cf-typegen:check, npm run typecheck, and npm run db:migrations:check, all of which succeeded locally.
• Attempted the full gate npm run test:ci; early checks completed but the full suite did not finish cleanly in this environment (long-running unrelated queue coverage runs surfaced and the run was not completed here).
• npm audit --audit-level=moderate could not complete in this environment due to the npm audit endpoint returning 403 Forbidden.

---

Codex Task

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.34%. Comparing base (db28d8f) to head (92b8641).
⚠️ Report is 5 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #4992       +/-   ##
===========================================
+ Coverage   74.88%   94.34%   +19.45%     
===========================================
  Files         473      473               
  Lines       39928    39932        +4     
  Branches    14557    14557               
===========================================
+ Hits        29902    37674     +7772     
+ Misses       8235     1583     -6652     
+ Partials     1791      675     -1116     

Flag	Coverage Δ
shard-1	46.34% <33.33%> (?)
shard-2	34.66% <0.00%> (?)
shard-3	32.03% <0.00%> (-0.01%)	⬇️
shard-4	31.95% <0.00%> (-0.01%)	⬇️
shard-5	33.76% <100.00%> (+<0.01%)	⬆️
shard-6	44.88% <0.00%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
src/db/schema.ts	73.07% <ø> (+1.09%)	⬆️
src/review/impact-map.ts	100.00% <100.00%> (ø)

... and 168 files with indirect coverage changes

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[loopover-orb]

Warning

🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨🟨

⏸️ Gittensory review result - manual review recommended

_{Review updated: 2026-07-12 08:35:44 UTC}

4 files · 1 AI reviewer · 1 blocker · readiness 100/100 · CI green · unstable

⏸️ Suggested Action - Manual Review

• No linked issue detected — If this PR is intended to solve an issue, link it explicitly in the PR body.

Review summary
This PR bounds unbounded growth of impact_map_query_cache by evicting expired rows on read and proactively deleting stale rows scoped to (project, repo) before each insert, adds a matching index in both the migration and Drizzle schema, and stops persisting the retrieved context blob (keeping only metrics_json). The core logic is sound and the two new regression tests genuinely exercise computeImpactMap end-to-end rather than fabricating state. The one real problem is that the index is added by editing the already-existing migrations/0132_impact_map_query_cache.sql file in place rather than via a new migration.

Blockers

• migrations/0132_impact_map_query_cache.sql:17-18 — this file already exists (the diff shows it as modified, not added) and its CREATE TABLE predates this PR per the 'fix(review): impact-map re-embeds + re-queries the vector index up to 20x per pass with no result cache #4500' comment already in the file, meaning it is very likely already applied to production D1; D1's migration ledger tracks applied state by filename, so appending `CREATE INDEX ...` to an already-applied file silently no-ops on existing deployments and only takes effect on fresh DBs, causing schema drift — add the index in a new file (e.g. migrations/0133_impact_map_query_cache_index.sql) instead of editing 0132.

Nits — 5 non-blocking

• src/review/impact-map.ts:118 — the new `DELETE ... WHERE fetched_at < ?` runs on every cache write (not just periodically), which is fine given the new index but is an extra round-trip per put; consider whether this should be probabilistic/periodic instead of unconditional if write volume is high.
• The PR doesn't cite an issue number it closes; per repo convention, contributor PRs should link the eligible open issue this addresses (the `fix(review): impact-map re-embeds + re-queries the vector index up to 20x per pass with no result cache #4500` reference is only a pre-existing code comment for the original cache table, not a link for this change).
• worker-configuration.d.ts carries a large unrelated regenerated diff (workerd bump, email/workflow API type changes) — expected from `cf-typegen` but worth calling out explicitly in the PR description as pure generated churn so reviewers don't need to diff it by hand.
• test/unit/impact-map.test.ts's cachingStorageStub DELETE-matching regex logic is a bit intricate (matchesExactRow vs matchesExpiredRepoRow) — a short comment explaining the two DELETE call shapes it's simulating would help future maintainers.
• Split the index addition into its own migration file per the schema-migrations convention so it can't be mistaken for editing an applied migration.

Concerns raised — review before merging

• No linked issue detected — If this PR is intended to solve an issue, link it explicitly in the PR body.

Signal	Result	Evidence
Code review	❌ 1 blocker	1 reviewer
Linked issue	✅ No-issue rationale	PR body explains why no issue is linked.
Related work	✅ No active overlap found	No same-issue or scoped active PR overlap found.
Change scope	✅ 20/20	Low review scope from cached public metadata (no linked issue context).
Validation posture	✅ 25/25	PR body includes validation/test evidence.
Contributor workload	✅ 10/10	Author activity: 45 registered-repo PR(s), 37 merged, 411 issue(s).
Contributor context	✅ Confirmed Gittensor contributor	JSONbored; Gittensor profile; 45 PR(s), 411 issue(s).
Gate result	❌ Blocking	Repo-configured hard blocker found.
Improvement	✅ Minor	risk: clean · value: minor — Code changes are accompanied by test evidence.

Review context

• Author: JSONbored
• Role context: owner (maintainer lane)
• Public audience mode: oss maintainer
• Lane context: Repository is configured for direct PR review.
• Public profile languages: not available
• Official Gittensor activity: 45 PR(s), 411 issue(s).
• PR-specific overlap: none found.

Contributor next steps

• Treat this as maintainer-lane context rather than normal contributor-lane activity.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Change scope = cached public metadata such as size labels, draft state, and review-burden hints.
• Validation posture = whether the PR provides enough public validation/test evidence for maintainer review.
• Contributor workload = public contributor activity and cleanup pressure, not a repo-wide quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

[BETA] Chat with Gittensory

Ask Gittensory a question about this PR directly in a comment — grounded only in the same cached, public-safe facts shown above, never a new claim.

• @gittensory ask <question> answers contribution-quality Q&A with source citations and freshness.
• @gittensory chat <question> answers in natural prose from cached decision-pack facts via local inference (maintainer/collaborator; read-only).
• A plain-language @gittensory mention with a real question is routed to the closest matching read-only command automatically -- no exact syntax required.

Full command reference: https://gittensory.aethereal.dev/docs/gittensory-commands

_{🟩 Safe / merged · 🟦 Advisory · 🟨 Held for review · 🟥 Blocked / closed}

---

💰 Earn for open-source contributions like this. Gittensor lets GitHub contributors earn for the work they already do — register to start earning →.

Checked by Gittensory, a quiet PR intelligence layer for OSS maintainers.

• Re-run Gittensory review