Agentic Actions Auditor
Audits GitHub Actions that use AI agents, especially the risk of untrusted input reaching the agent.
How to use it
Install it with the open Skills CLI (needs Node.js), then restart your agent:
npx skills add trailofbits/skills --skill agentic-actions-auditor- Download the ZIP from this page and unzip it.
- Put the
agentic-actions-auditorfolder in.claude/skills/inside your project (or~/.claude/skills/for all projects). - Restart Claude Code and call it with
/agentic-actions-auditor.
بالعربي
يراجع GitHub Actions التي تستخدم وكلاء ذكاء اصطناعي، خصوصا خطر وصول مدخلات غير موثوقة إلى الوكيل.
More for security
- /security-review (built into Claude Code): Runs a full security review of the pending changes on your branch.
- Security Review: Reviews auth, input handling, secrets and APIs before release, using a practical security checklist.
- Audit Context Building: Builds a precise picture of code flow, assumptions and dependencies before a security audit, so findings have context.