Skip to main content
affpapa.org
RU EN
skill Claude Code CX Codex Cursor GitHub Copilot

agentic-actions-auditor — Claude Code skill: setup and audit

Inspects GitHub Actions workflows for prompt injection risks and insecure AI agent setups. Catches attacker inputs leaking into prompts via event contexts or environment variables.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓ 
npx skills add trailofbits/skills --skill agentic-actions-auditor -a claude-code
Open on GitHub

What agentic-actions-auditor does

This skill provides static inspection rules for GitHub Actions workflows that run Claude Code Action or Gemini CLI. It checks whether untrusted input reaches agent prompts directly or passes into prompt parameters through env: definitions. Dangerous triggers like pull_request_target and issue_comment receive special attention because external contributors can invoke them without maintainer rights.

The audit runs against local workflow YAML files or remote repositories via gh api. It inspects sandbox flags and wildcard tool permissions without executing remote code or mutating workflows. It handles static discovery rather than runtime prompt injection fuzzing.

What it helps with

  • detect untrusted input reaching AI prompts via environment variables
  • identify risky pull_request_target triggers in agentic pipelines
  • spot disabled sandbox flags in workflow steps
  • audit remote GitHub repositories for AI action vulnerabilities

Example prompts for the agent

  • Audit our GitHub Actions workflows for AI prompt injection risks
  • Analyze repository owner/repo for vulnerable agentic CI configurations

How to install agentic-actions-auditor

1Skills CLI (Vercel)

npx skills add trailofbits/skills --skill agentic-actions-auditor -a claude-code

2Manually: put the files in place

~/.claude/skills/agentic-actions-auditor/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/agentic-actions-auditor/

From the same set35

The whole set →
semgrep trailofbits Runs a Semgrep static security audit on your codebase. It presents a plan for user confirmation and merges all findings into SARIF. Claude Code CX Codex Cursor +1 9,500installs differential-review trailofbits Reviews code changes for security issues using Git history and test coverage. Use it for commits, diffs, and pull requests that need a written review report. Claude Code CX Codex Cursor +1 8,800installs modern-python trailofbits Configures Python projects with a modern tooling stack: uv for dependency management, ruff for linting and formatting, ty for type checking. Use when starting new projects, writing scripts, or migrating from legacy tools. Claude Code CX Codex Cursor +1 8,700installs secure-workflow-guide trailofbits Guides smart contracts through Trail of Bits' 5-step security audit on check-in or before deployment. It runs static analysis, generates visual charts, and prepares fuzzing tests. Claude Code CX Codex Cursor +1 7,400installs sharp-edges trailofbits Audits API designs and configuration schemas for developer footguns. Flags dangerous defaults and fragile interfaces where the easiest path causes security vulnerabilities. Claude Code CX Codex Cursor +1 6,900installs gh-cli trailofbits A skill that redirects GitHub requests from curl and fetch to authenticated gh CLI for repositories, pull requests, and issues. Use it for private repos and higher API limits. Claude Code CX Codex Cursor +1 6,600installs variant-analysis trailofbits Finds related occurrences of an already identified defect across a repository. Use it after a vulnerability or logic issue turns up and similar code needs checking. Claude Code CX Codex Cursor +1 6,400installs sarif-parsing trailofbits Processes existing SARIF reports from CodeQL, Semgrep, and other scanners. Use it to filter findings, remove duplicates, or prepare results for CI/CD. Claude Code CX Codex Cursor +1 5,900installs

Similar

the-fool https://github.com/Jeffallan Challenges plans, architectures, and decisions using five structured reasoning modes. Useful for spotting blind spots and failure modes before writing code. Claude Code CX Codex Cursor +1 4,400installs security-review getsentry Finds injections, XSS, authorization flaws, and other vulnerabilities in code. Reports only high-confidence findings. Claude Code CX Codex Cursor +1 21,100installs gateway-workers cloudflare Skill for building, debugging, and reviewing Cloudflare Workers triggered by Forward actions in Gateway HTTP policies. Work with GatewayMiddleware, context types, and request continuation. Claude Code CX Codex Cursor +1 7,700installs agentic-jujutsu ruvnet Coordinates multiple AI agents working on code simultaneously without version control locks, learning from execution trajectories. Claude Code CX Codex Cursor +1 1,200installs git-guardrails-claude-code Matt Pocock This skill adds a Claude Code hook that stops risky Git commands before they run. Use it when an agent must not push or discard working-tree changes. Claude Code 499,069installs ponytail-review DietrichGebert ponytail-review reviews a change beyond the diff, checking connected code, bugs, risks, load, tests, speed, and unnecessary code. Use it for PRs and commits. Claude Code CX Codex Cursor +1 56,300installs ponytail-audit DietrichGebert It audits a repository and ranks bugs, risks, and redundant code in a one-shot report. Use it when you need a review without code changes. Claude Code CX Codex Cursor +1 44,800installs security-audit cloudflare Reviews codebases and APIs for vulnerabilities using safe sandboxing. Fits pre-release security reviews, pen-test triage, and architecture checks. Claude Code CX Codex Cursor +1 52,200installs

Questions about agentic-actions-auditor

How do I install agentic-actions-auditor in Claude Code?

Run npx skills add trailofbits/skills --skill agentic-actions-auditor -a claude-code.

What does agentic-actions-auditor do?

It audits GitHub Actions workflow files to locate attack paths where untrusted external data reaches AI coding agents.

Does the skill automatically modify or fix vulnerable workflow files?

No, it strictly reports security findings and leaves file modifications to the developer.

See also

Project sponsors

Companies that keep this analytics open
Gold sponsor
Nutra.Club

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor