Trail of Bits Agentic Actions Auditor
Find prompt injection risks in GitHub Actions that run Claude or other AI agents.
Official vendorby Trail of BitsChecked October 18, 2026
A Trail of Bits skill for a new class of risk: AI coding agents running in CI. It audits workflows that call agents such as Claude Code Action and finds paths where attacker-controlled text can reach them.
Who it is for
- Security engineers
- DevOps engineers
- Open source maintainers
Prompts to try
Once the skill is installed, ask for the result in plain words. Claude picks the skill up on its own.
Audit our GitHub workflows that use Claude Code ActionCan an outside contributor prompt-inject our CI agent?Review this repo's AI-powered actions for security problemsHow to install it
In Claude Code
/plugin marketplace add trailofbits/skills/plugin install agentic-actions-auditor@trailofbits
In Claude.ai
Not practical in Claude.ai because the skill drives local command-line tools, so use Claude Code.
Good to know
Can also audit workflows from a GitHub URL.
Unlock the install steps
Free. Leave your email once to open the install steps for all 149 skills, and get the best new skills each month.
How we checked it
Confirmed SKILL.md at plugins/agentic-actions-auditor/skills/agentic-actions-auditor/SKILL.md in trailofbits/skills via the GitHub API; skill folder last changed 2026-09-16, repo has 7,620 stars and was pushed 2026-10-15.