feat(enrichment): secrets-in-logs & PII-egress scanner

[oktofeesh1]

Closes #1507.

Adds a REES analyzer that flags added lines passing sensitive data into a logging or stdout sink — console.log(req.headers.authorization), logger.info(\token=${apiKey}`), console.log(req)`. This is distinct from the shipped hardcoded-secret scan: that inspects literal values; this inspects the data flow into a sink (a secret reaching a log is a leak even when the value isn't a literal). It's exactly the kind of pass the no-checkout in-prompt reviewer can't reliably do.

Approach

Pure compute, no network, no new dependency (a structural detector, not a bundled Semgrep). Precision-first so it doesn't generate noise:

• A linear codeOnly() hand-scan blanks string-literal messages (keeping ${…} interpolation bodies, which are real code) before matching — so console.log("password reset") is not flagged.
• A hit requires a sensitive name used as code: a property access (.authorization), a ${…} interpolation (${apiKey}), or a dumped request/session object (req, req.headers/body/cookies/session). Innocuous request scalars (req.method/url/path) are excluded.
• All matchers are flat, linear-time regexes (no nested-quantifier shapes) — the analyzer can't be the DoS class it sits beside.
• Reports location + sink + category only; never the logged value. Line-cited via hunk headers; findings bounded.

Files (all inside review-enrichment/, per the established analyzer pattern)

• src/types.ts — SecretLogFinding type + secretLog BriefFindings key
• src/analyzers/secret-log.ts — the analyzer (codeOnly + detectSecretLog + scanSecretLog)
• src/brief.ts — registered in ANALYZERS
• src/render.ts — public-safe brief block
• test/enrichment.test.ts — codeOnly, classifier (incl. false-positive guards), line-cited scan, render, and a buildBrief integration test

Validation

• npm test (build + node --test) inside review-enrichment/: 50/50 pass (6 new).
• git diff --check clean; prettier-clean; engine src/** untouched (outside the engine tsc/vitest/codecov scope — zero conflict).

[superagent-security]

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

[loopover-orb]

Tip

🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩

✅ Gittensory review — safe to merge

5 files · 1 AI reviewers · no blockers · readiness 55/100 · CI green · clean

✅ Approved — safe to merge

Review summary
Adds a secrets-in-logs / PII-egress analyzer that correctly distinguishes sensitive-value data flow into sinks from string literals that merely mention sensitive words. The `codeOnly` linear pass is well-implemented — the escape-aware single/double-quoted branch and the depth-tracking `${…}` interpolation keeper are both correct. Hunk-header line-citation matches the established pattern; the `ANALYZERS` registry wiring, `BriefFindings` key, and `renderBrief` block are all consistent with prior analyzers. No blocking defects found.

Signal	Result	Evidence
Code review	✅ No blockers	1 reviewers, synthesized
Linked issue	✅ Linked	#1507
Related work	⚠️ 3 scoped overlaps	Top overlaps are listed below; lower-confidence bulk is hidden.
Review load	❌ 8/20	Readiness component derived from cached public PR metadata and labels; size label size:L.
Validation evidence	❌ 5/25	Cached preflight status is hold.
Open PR queue	❌ 3/10	33 open PR(s), 17 likely reviewable, 16 unlinked.
Contributor context	✅ Confirmed Gittensor contributor	oktofeesh1; Gittensor profile; 668 PR(s), 7 issue(s).
Gate result	✅ Passing	No configured blocker found.

Nits — 5 non-blocking

• secret-log.ts: The second arm of `REQUEST_OBJECT_RE` — `\b(?:headers|session|cookies)\s*(?:\)|\.[\s\S]*?[\w$])` — flags ANY property access on a standalone `headers`/`session`/`cookies` variable (e.g., `logger.info(headers.contentType)` or `logger.info(cookies.preferredLanguage)`), not only dumps of the full object. The PR's benign-field exclusion logic covers only `req.method`/`req.url`/`req.path`, leaving the second arm notably broader than documented; add a comment acknowledging this false-positive surface or tighten the pattern to enumerate safe sub-properties the same way the `req.*` arm does.
• enrichment.test.ts: The `process.stdout.write(session.cookie)` assertion only checks `.sink`, leaving the `.category` return path (which resolves to `request-object` via the second `REQUEST_OBJECT_RE` arm) entirely unverified — add `assert.equal(...?.category, 'request-object')` on the same line.
• secret-log.ts (scanPatchForSecretLog): The `body.length > MAX_LINE_CHARS` skip branch has no test — add a test patch line of >2000 chars and assert it produces no finding.
• enrichment.test.ts (scanSecretLog): The 25-finding cap (`MAX_FINDINGS`) in `scanSecretLog` is untested — the existing test sends exactly one finding. Add a test that drives 26+ findings across files and asserts exactly 25 are returned.
• secret-log.ts: Added diff lines that are inline comments (`// console.log(req.headers.authorization)`) will be flagged since no comment-prefix filter is applied before `SINK_RE`; this is a known limitation of line-level structural scanning, but it is undocumented — add a comment or a TODO so future contributors don't mistake it for an oversight.

Review context

• Author: oktofeesh1
• Role context: outside_contributor
• Public audience mode: oss maintainer
• Lane context: Repository registration is not available in the local Gittensory cache.
• Public profile languages: JavaScript, MDX, Python, TypeScript
• Official Gittensor activity: 668 PR(s), 7 issue(s).
• Related work: Items reference the same linked issue feat(enrichment): Secrets-in-logs & PII-egress scanner #1507. (issue #1507)
• Related work: Open PR work references issue feat(enrichment): Secrets-in-logs & PII-egress scanner #1507. (issue #1507)
• Related work: Items reference the same linked issue MCP CLI: add PowerShell shell completion #1452. (issue #1452, PR #1448)
• Additional title-only matches omitted; title-only overlap does not block.

Contributor next steps

• Review top overlaps.
• Add scope summary.
• Fix blocker.
• Expect slower review.
• Refresh registry data or choose a registered active repo.
• Check active issues and PRs before submitting.

Signal definitions

• Related work = same linked issue, overlapping active PRs, or title/path similarity.
• Review load = cached public PR metadata such as size labels, changed paths, and preflight status.
• Open PR queue = repo-wide review pressure; it is not a PR quality failure.
• Contributor context = public GitHub/Gittensor identity context; non-Gittensor status is not a blocker.

Review details

Generated from public PR metadata and the diff. Advisory only; deterministic signals remain authoritative.

Adds a secrets-in-logs / PII-egress analyzer that correctly distinguishes sensitive-value data flow into sinks from string literals that merely mention sensitive words. The `codeOnly` linear pass is well-implemented — the escape-aware single/double-quoted branch and the depth-tracking `${…}` interpolation keeper are both correct. Hunk-header line-citation matches the established pattern; the `ANALYZERS` registry wiring, `BriefFindings` key, and `renderBrief` block are all consistent with prior analyzers. No blocking defects found.

Nits (5)

• secret-log.ts: The second arm of `REQUEST_OBJECT_RE` — `\b(?:headers|session|cookies)\s*(?:\)|\.[\s\S]*?[\w$])` — flags ANY property access on a standalone `headers`/`session`/`cookies` variable (e.g., `logger.info(headers.contentType)` or `logger.info(cookies.preferredLanguage)`), not only dumps of the full object. The PR's benign-field exclusion logic covers only `req.method`/`req.url`/`req.path`, leaving the second arm notably broader than documented; add a comment acknowledging this false-positive surface or tighten the pattern to enumerate safe sub-properties the same way the `req.*` arm does.
• enrichment.test.ts: The `process.stdout.write(session.cookie)` assertion only checks `.sink`, leaving the `.category` return path (which resolves to `request-object` via the second `REQUEST_OBJECT_RE` arm) entirely unverified — add `assert.equal(...?.category, 'request-object')` on the same line.
• secret-log.ts (scanPatchForSecretLog): The `body.length > MAX_LINE_CHARS` skip branch has no test — add a test patch line of >2000 chars and assert it produces no finding.
• enrichment.test.ts (scanSecretLog): The 25-finding cap (`MAX_FINDINGS`) in `scanSecretLog` is untested — the existing test sends exactly one finding. Add a test that drives 26+ findings across files and asserts exactly 25 are returned.
• secret-log.ts: Added diff lines that are inline comments (`// console.log(req.headers.authorization)`) will be flagged since no comment-prefix filter is applied before `SINK_RE`; this is a known limitation of line-level structural scanning, but it is undocumented — add a comment or a TODO so future contributors don't mistake it for an oversight.

_{🟩 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