alexgreensh/repo-forensics / skills/repo-forensics
repo-forensics
FlaggedSecurity forensics for git repos, AI skills, and MCP servers. Audits dependencies, detects prompt injection, credential theft, runtime dynamism, manifest drift, known CVEs, CISA KEV (actively exploited) vulns, and 2026 attack patterns. Not for fixing vulnerabilities or pentesting.
tests/corpus/DEAD_ANCHORS_CORPUS.md
View on GitHub# Dead-Anchor Test Corpus — status & TODO
Fixtures backing `scan_dead_anchors.py`. All unit/integration tests monkeypatch
the network (vuln_feed's `urllib.request.urlopen` chokepoint + `socket.gethostbyname`);
NO test in `tests/test_dead_anchors_*.py` ever performs a real HTTP request or
DNS lookup.
## Committed benign fixtures (in `benign/`)
- `dead_anchors_live_refs.md` — live GitHub owner/repo + `github:` shorthand,
live npm (`lodash`) / PyPI (`requests`) install commands, safe-allowlisted docs
domains. Asserted to produce ZERO findings under the LO branch
(`test_scan_dead_anchors.py::test_benign_corpus_all_lo_zero_findings`, network
mocked to 200/registered). Budgeted at zero for the per-file committed-corpus
gate too (see `budgets.json`).
## Positive (teeth) coverage
Lives in `tests/test_scan_dead_anchors.py` and `tests/test_dead_anchors_probe.py`
(canned 404 / NXDOMAIN / provider-fingerprint bodies per anchor type → correct
CRITICAL/HIGH/MEDIUM finding). Negative coverage: canned 200/registered → zero
findings; canned 403/429/5xx/timeout → zero findings + no crash (never-hard-fail).
## TODO — external datasets (MANUAL, security-gated — do NOT automate)
These are deliberately NOT wired into CI yet. Each requires a manual, blocking
`repo-forensics`-first security review before any external bytes enter this
corpus (Alex's standing external-code rule applied to the test data itself).
1. **SkillSieve** (`github.com/xiaohou521/skillsieve`, 49,592 skills + 400
labeled). Before copying ANY labeled skill folder into `tests/corpus/`: clone
into an isolated temp dir, run a full `repo-forensics` scan on the clone,
review findings, and only then extract reviewed folders. Do NOT clone/fetch
from a test or CI path.
2. **ToxicSkills 8 live-malicious IOC URLs** (Snyk, e.g. `clawhub.ai/zaycv/clawhud`).
These belong in `ioc_manager.py`'s SIGNED remote IOC feed (a feed-authoring
step publishing to the `iocs/latest.json` the manager pulls — NOT a direct
edit to any file under `scripts/`/`data/`). Once fed, add a negative-control
test asserting `scan_dead_anchors` does not double-fire or conflict with
`ioc_manager`'s existing IOC flag on the same anchor (IOC-badness ≠
claimability; distinguishable by `scanner` field + `category`).
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |