SuperAPI AI Directory

endpoint-pentest

Testing & quality

View repository

Updated 10-16 08:38

About this skill

Endpoint / API Pentest

Authorized, safe-by-default HTTP/API testing. Read ../../SECURITY_RULES.md first. Conservative rates; owned/staging targets; read-only unless a write is explicitly authorized against a disposable record.

Deliverables

  • Route inventory
  • Endpoint Risk Matrix (../../templates/endpoint-risk-matrix.md)
  • Confirmed / Suspected Findings, Passed Checks
  • Fix plan + regression tests

Phases

  1. Scope & authorization.
  2. Route discovery — from OpenAPI/Swagger, JS bundles, run-httpx.sh, run-ffuf-safe.sh (owned/staging only), sitemap, robots.
  3. Spec leakage — is OpenAPI/Swagger/GraphQL introspection public? Does it leak internal routes?
  4. AuthN — token validation negative tests (deny expected).
  5. AuthZ — IDOR/BOLA (object level) + broken function-level authorization (admin routes as normal user).
  6. Mass assignment — extra fields (is_admin, role, owner_id) accepted?
  7. Excessive data exposure — responses leaking fields the UI hides.
  8. SSRF — URL/webhook/import params; ../../payloads/ssrf.txt.
  9. Injection — SQL/NoSQL/command/template, safe probes only.
  10. File upload — type/size/content-type validation; path of stored file.
  11. Webhook signature bypass — missing/forged/replayed signatures.
  12. CORS — wildcard + credentials; reflected origin.
  13. CSRF — where cookies authenticate state-changing routes.
  14. Rate limiting & pagination abuse — realistic abuse, low volume.
  15. Cache poisoning — unkeyed headers influencing cached responses.
  16. Error/debug leakage — stack traces, debug endpoints, health endpoints leaking versions/config, admin endpoints exposed.
  17. Request size / timeout — limits enforced.
  18. Sensitive-data logging — tokens/PII in logs.

Full checklist

endpoint-checklist.md. Payloads: payload-catalog.md + ../../payloads/.

Report

report-template.md or hand to security-report-writer.