agentic-actions-auditor
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is purely defensive and educational. It provides comprehensive documentation on identifying common attack vectors in GitHub Actions AI integrations, such as direct expression injection and environment variable intermediaries.
- [SAFE]: All tool usage is restricted to analytical tasks. The Bash instructions are scoped to listing and fetching repository content via the GitHub CLI (
gh api), which is consistent with the tool's auditing purpose. - [SAFE]: The methodology includes explicit safety rules, such as prohibiting the execution of fetched YAML content and treating remote data strictly as read-only material for analysis.
- [SAFE]: No instances of obfuscation, hidden URLs, or unauthorized data access were detected. The references to external repositories are provided as educational examples of vulnerable patterns for auditing purposes.
Audit Metadata