verify-before-filing
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a structured decision-making framework to prevent unnecessary or incorrect outbound communication to external repositories.
- [COMMAND_EXECUTION]: Shell commands are restricted to informational queries using the GitHub CLI (
gh search) for searching existing issues and pull requests. This is a legitimate use case for the stated purpose. - [DATA_EXPOSURE]: No hardcoded credentials or sensitive file paths were detected. The skill uses tools like
fetchandsourcegraphto read public source code for verification purposes. - [PROMPT_INJECTION]: The skill instructions do not attempt to bypass agent safety filters or override system prompts. The triggers are functionally relevant to activating the verification workflow.
Audit Metadata