ci-first-verification
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use the GitHub CLI (gh) to check PR status and workflow runs (e.g., gh pr checks, gh run list). This is a legitimate use of developer tools to manage resources.
- [PROMPT_INJECTION]: The skill directs the agent to read failure logs from external CI runs (gh run view --log-failed). This constitutes an indirect prompt injection surface as logs contain data from external processes. * Ingestion points: CI log output retrieved via the GitHub CLI in SKILL.md. * Boundary markers: The instructions do not define specific delimiters or 'ignore' warnings for the log content. * Capability inventory: The agent maintains standard shell execution and file system access capabilities. * Sanitization: There is no mention of filtering, escaping, or sanitizing the log output before the agent processes it.
Audit Metadata