agentic-actions-auditor
Pass
Audited by Gen Agent Trust Hub on Jul 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill facilitates remote auditing by fetching workflow files from GitHub via the official API. This process targets a well-known service and is limited to the retrieval of text-based configuration for static analysis.
- [COMMAND_EXECUTION]: Shell usage is restricted to the official GitHub CLI (
gh) for listing and reading files. The instructions provide explicit safety guardrails that forbid the execution of any retrieved data, preventing potential command injection or remote code execution. - [PROMPT_INJECTION]: The skill handles untrusted workflow content which could contain injection payloads. To mitigate this risk, it incorporates educational framing that instructs the agent to treat external content as data only and maintain a critical security posture regardless of developer rationalizations.
Audit Metadata