ajtazer/heckit / skills/cve-hunting/cve-hunting-methodology
cve-hunting-methodology
UnverifiedCross-cutting methodology for the CVE-hunting pipeline (ported from find-cve-agent): when a bug is design-vs-real, false-positive avoidance, version checking before reporting, responsible disclosure norms, acceptance-rate expectations per vuln class, vulnerability chaining playbook, known false-positive patterns, maintainer-response patterns, secure-pattern reference (what NOT to flag), and a self-criticism checklist to run before claiming a finding. Use this alongside the cve-hunter/cve-exploiter/cve-validator agents and the /cve-hunt pipeline — load it whenever judging if a finding is real, worth reporting, or already fixed.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 2.7 KB |
| references/acceptance-rates.md | markdown | 12.1 KB |
| references/chaining-playbook.md | markdown | 10.3 KB |
| references/design-vs-bug.md | markdown | 3.6 KB |
| references/disclosure-channels.md | markdown | 11.0 KB |
| references/false-positive-avoidance.md | markdown | 4.0 KB |
| references/false-positive-patterns.md | markdown | 12.0 KB |
| references/maintainer-responses.md | markdown | 11.6 KB |
| references/rejection-taxonomy.md | markdown | 12.0 KB |
| references/responsible-disclosure.md | markdown | 2.4 KB |
| references/secure-patterns.md | markdown | 15.0 KB |
| references/self-criticism-checklist.md | markdown | 10.2 KB |
| references/version-checking.md | markdown | 2.4 KB |
references/false-positive-avoidance.md
View on GitHub# False Positive Avoidance — Mandatory Pre-Submission Checks
**Every finding MUST pass ALL of these checks before submission. No exceptions.**
## Gate 1: fp-check Verification (6 Gates)
Run the fp-check skill. All 6 gates must pass:
| Gate | Question | FAIL means |
|------|----------|------------|
| 1. Process | Were all analysis phases completed with evidence? | Incomplete analysis |
| 2. Reachability | Can an attacker actually reach and control data at the vulnerable point? | Theoretical only |
| 3. Real Impact | Does exploitation lead to real security consequences? | No meaningful impact |
| 4. PoC Validation | Does the PoC demonstrate the actual attack path? | Unproven claim |
| 5. Math Bounds | Does mathematical analysis confirm the vulnerable condition? | Numbers don't add up |
| 6. Environment | Do environmental protections entirely prevent exploitation? | Blocked at runtime |
## Gate 2: Self-Criticism Checklist (7 Items)
Answer honestly. If ANY answer is "yes", investigate further before submitting:
1. **README warning?** — Does the README/docs warn against untrusted input?
2. **Intended behavior?** — Is this documented/designed behavior, not a bug?
3. **Already handled?** — Does the library already handle this gracefully?
4. **Alpha/beta?** — Is this pre-release? Will the maintainer issue a CVE?
5. **Over-researched?** — How many recent CVEs? Am I racing other researchers?
6. **JSON.parse equivalence?** — For clobbering: does JSON.parse do the same thing? Can I show a REAL crash?
7. **Error severity?** — For recursion: is it OOM (real DoS) or just RangeError (caught)?
## Gate 3: Version Verification
1. Read `package.json` / `requirements.txt` / `go.mod` for exact version
2. Search NVD for that exact version — may already be patched
3. Search GHSA and OSV.dev for existing advisories
4. Check CHANGELOG for "security fix" entries
5. Check git log for security-related commits since that version
6. **Test against the LATEST release**, not just whatever is in node_modules
## Gate 4: Default Configuration Test
- Only report vulnerabilities that exist in the **default configuration**
- If a non-default option must be enabled, the severity drops significantly
- Document any configuration requirements in the report
## Gate 5: PoC Execution (3x Minimum)
1. Run PoC against clean install of latest version
2. Run 3 times minimum — must succeed every time
3. **Fail 3 times = FALSE POSITIVE. Move on immediately. No exceptions.**
4. Evidence must match the claimed impact exactly
## Gate 6: Similar CVE Search
Before submitting, search all of:
- NVD (nvd.nist.gov) — `scripts/check-nvd.sh <package>`
- GitHub Advisory Database (github.com/advisories)
- OSV.dev — `scripts/check-osv.sh <ecosystem> <package>`
- Snyk vulnerability database (snyk.io/vuln)
If a similar CVE exists, determine:
- Is this the same issue? → DUPLICATE, do not submit
- Is this an incomplete fix of a prior CVE? → Document the prior CVE and explain what's still broken
- Is this a different vulnerability in the same component? → Proceed, reference the prior CVE
## Historical False Positives to Learn From
| Project | Claimed | Reality | Lesson |
|---------|---------|---------|--------|
| FUXA | RCE via npm install scripts | live-plugin-manager doesn't run install scripts | Package download != code execution |
| Rundeck | Path traversal file read | Admins can already read files via exec | Check if access is genuinely new |
| Strapi | SSRF via IPv6 bypass | v2.0.0 blocks ALL localhost variants | Test the actual library version |
| Payload CMS | CRLF header injection | Node.js rejects CRLF at runtime level | Check runtime protections |
| Flowise | NodeVM sandbox escape | @flowiseai/nodevm blocks constructor chains | Test actual implementation |
| sanitize-html | XSS bypass | Browser XSS protection, not library bug | Know library vs browser boundary |
| validator.js | ReDoS | No exponential backtracking confirmed | Measure actual regex growth rate |
| qs v6.14.2 | Prototype pollution | Already patched in this exact version | Always check exact version number |
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: 194k | Updated | Claude CodeCodex | Verified | Install |