seed-snapshot-release
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes multiple shell commands using the GitHub CLI and git, including git push and gh pr comment, which modify the remote repository state.
- [INDIRECT_PROMPT_INJECTION]: The skill reads GitHub PR comments to determine triggering events and to report results to the user. This creates a surface where an attacker could inject malicious content into PR comments to influence the agent's behavior. Ingestion: gh pr view in SKILL.md. Boundaries: Absent. Capabilities: git push, gh pr comment, gh run watch. Sanitization: jq extraction.
- [DYNAMIC_EXECUTION]: The skill assembles a bash script at runtime using string substitution for variables like PR numbers and timestamps. While the template is static, dynamic command assembly is a known risk factor. Evidence: Script template in Step 3 of SKILL.md.
Audit Metadata