Cloudflare Open-Sources Security-Audit Skill for Independently Verified AI Code Scans
Summary
Cloudflare open-sources a security-audit skill that guides coding agents through six audit phases and emits independently verified, machine-readable vulnerability findings, requiring confirmed bugs to include a complete source trace and bounded observed result while sandboxing target code with networking disabled.
Key Points
- Cloudflare releases the open-source security-audit skill, which orchestrates coding agents through six audit phases and produces independently verified, machine-readable vulnerability findings.
- The skill classifies findings as confirmed, needs_validation, or rejected; confirmed findings require a complete source trace and bounded observed result, while unresolved leads receive no severity.
- Full audits default their output to ~/security-audit-skill/<repo-name>/run-<N>, and target-controlled code runs only in an OS-enforced sandbox with networking disabled and resource limits.