ship
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the local file system.\n
- Ingestion points: Step 1 in SKILL.md requires running git status and reading changed files.\n
- Boundary markers: No delimiters or 'ignore' instructions are defined for the file content.\n
- Capability inventory: Includes git operations (commit, push, merge, checkout), pull request creation, and issue management.\n
- Sanitization: Content is read and analyzed without specific sanitization or escaping mechanisms.\n- [COMMAND_EXECUTION]: The skill relies on shell-based git commands to perform repository management and monitoring tasks.
Audit Metadata