Skip to content

Security research backlog (offensive + CWE)

Status: Active
Agent: security_auditor
Goal: offensive_security
Branch: cursor/security-research-loop

todos:

  • id: sec-r0-cwe-delta content: "CWE Top 25 feed sync vs cve-catalog.json — map new CWEs to catalog_gaps and li-tests" status: completed study_only: true

  • id: sec-r1-httpd-fuzz-smoke content: "security-research: pending plan todo sec-r1-httpd-fuzz-smoke — gap orchestrator" status: pending study_only: true

  • id: sec-r2-tier5-gap-exploit content: "security-research: pending plan todo sec-r2-tier5-gap-exploit — gap orchestrator" status: pending

  • id: sec-r3-runtime-surface content: "security-research: pending plan todo sec-r3-runtime-surface — gap orchestrator" status: pending

Agent instructions

  • One todo per loop iteration (security-research-plan-loop.py).
  • Agent: security_auditor (LI_SECURITY_PLAN_AGENT).
  • Gates: ./scripts/security-research-gates.sh.
  • Deliverable: docs/security/studies/YYYY-MM-DD-<todo-id>.md (see security-research-grading.md).
  • Survey todos (study_only: true): gates require study file; benches optional unless code changes.
  • Push branch cursor/security-research-loop every iteration.