code-review
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple sources including git diff outputs, git log commit messages, and external issue tracker content. This data is interpolated into prompts for parallel sub-agents without clear boundary markers or sanitization.
- Ingestion points: Git diff hunks, commit messages (via git log), and issue tracker implementation contracts fetched from external sources (SKILL.md).
- Boundary markers: Absent. The instructions do not specify delimiters or instructions to ignore embedded commands within the diff or log data.
- Capability inventory: The skill has the capability to spawn parallel sub-agents and read local repository files (e.g., CODING_STANDARDS.md, docs/).
- Sanitization: Absent. There is no mention of filtering or escaping content from the git history or external issues before processing.
- [COMMAND_EXECUTION]: The skill relies on executing local system commands such as git diff, git log, and git rev-parse to function. While these are necessary for code review, they provide a vector for potential command injection if the 'fixed point' supplied by a user is not strictly validated by the underlying platform tool.
Audit Metadata