The Agent Skills Directory

[SAFE]: The skill's primary purpose is a security-focused code review and remediation tool. It includes explicit layers for SAST (Source Analysis), Secrets detection, and SCA (Software Composition Analysis) to protect the user's codebase.
[SAFE]: The skill uses a 'read-only' posture for its analysis subagents, ensuring that automated agents cannot modify the codebase until the final remediation phase, which includes its own 'Scope-Guard' to prevent unintended large-scale changes.
[SAFE]: Data exfiltration is mitigated by the fact that the skill focuses on local code analysis and uses the web-researcher role only for general best-practices research, not for sending private code to external domains.
[SAFE]: The remediation phase includes a mandatory validation step (Phase 4.5) with confidence scoring to ensure only high-confidence, verified findings are addressed, reducing the risk of introducing automated bugs.
[SAFE]: The skill explicitly forbids the use of commit or push commands, ensuring the user retains final control over all changes made to the codebase.

review-story