chore: final audit pass — accuracy, SVGs, CLI parity#328
Merged
andres-linero merged 3 commits intomainfrom Mar 8, 2026
Merged
Conversation
ResponseInspector already checks RESPONSE_INJECTION_PATTERNS for all tools. VectorDBInjectionDetector.check() also always runs injection patterns, so calling both for non-vector tools produced duplicate alerts. Fix: only call vector_detector.check() when is_vector_tool() returns True. Non-vector tools: ResponseInspector covers cloaking + SVG + invisible + injection. Vector tools: ResponseInspector (HIGH) + VectorDBInjectionDetector (CRITICAL, cache_poison_* tagged) — intentional dual signal for elevated risk.
ARCHITECTURE.md: - auth.py description: JWT (wrong) → scrypt KDF + RBAC API keys (correct) - Runtime detectors: 6 → 7 (VectorDBInjectionDetector) - enforcement.py: 10 checks → 8 (counted actual def check_* functions) - CLI commands: 15+ → 22+ - audit_log.py description updated: HMAC-SHA256 signed SVGs (7 files, dark + light variants): - engine-internals: "6 detectors" → "7 detectors (drift, injection, credential leak, vector DB)" - modes-flow: "6 detectors: drift, injection, cred leak" → "7 detectors" - offerings-map: "6 detectors + alert webhook" → "7 detectors + alert webhook" - scanner-architecture: "6 detectors" → "7 detectors" CLI (cli.py): - api command: add --log-level and --log-json flags (matching serve command) - Wire setup_logging() into api command body - Pass log_level to uvicorn (was hardcoded "info") AUDIT.md: - Add Areas 7-9 (CLI, discovery/peripheral, accuracy) - Final summary table covering all 9 audit areas - 13 bugs fixed total (F1-F13), 3 open issues remain (O1-O3, O5)
Contributor
Dependency Review✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.Scanned FilesNone |
andres-linero
approved these changes
Mar 8, 2026
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
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
Final pass of the 9-area codebase audit. All 7 criteria fully covered.
Changes
ARCHITECTURE.md — 5 inaccuracies fixed
auth.py — JWT authentication→ scrypt KDF API keys, RBAC roles (no JWT anywhere in codebase)Runtime detectors: 6→ 7 (VectorDBInjectionDetector added in fix: wire VectorDBInjectionDetector into proxy + runtime audit #326)VectorDBInjectionDetectorCLI with 15+ commands→ 22+ commands and groupsenforcement.py — 10 checks→ 8 checks (verified by countingdef check_*functions)SVGs — 7 files updated
"6 detectors"→"7 detectors"in engine-internals, modes-flow, offerings-map, scanner-architecture (dark + light variants)CLI (
apicommand)--log-leveland--log-jsonflags (matching theservecommand — consistency fix)setup_logging()now called at api command startupuvicorn log_levelwired to user input instead of hardcoded"info"AUDIT.md
Test plan
tests/test_api_endpoints.py+tests/test_api_hardening.py— 25 passed