change-description
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the GitHub CLI (
gh) to perform repository operations.\n - Evidence: The process concludes with the command
gh pr edit --body-fileorgh pr create(SKILL.md, Process Step 9).\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from external sources, which could contain malicious instructions.\n - Ingestion points: The agent reads Git diffs and existing pull request conversations (SKILL.md, Process Step 2).\n
- Boundary markers: No delimiters or instructions are provided to the agent to disregard potential instructions in the ingested data.\n
- Capability inventory: The agent has write access to repository metadata via
ghCLI commands executed autonomously (SKILL.md, Process Step 9).\n - Sanitization: No sanitization is performed on the ingested content.\n- [PROMPT_INJECTION]: The skill instructs the agent to bypass standard operational safeguards.\n
- Evidence: Step 9 in the process explicitly tells the agent to publish the pull request description 'without asking', which bypasses the user confirmation step usually required for impactful repository actions.
Audit Metadata