code-review
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the git repository and external issue trackers, which creates an attack surface where malicious instructions could be embedded in commit messages or documentation.
- Ingestion points: The skill fetches content from commit messages (Step 2.1), external issue trackers, and repository files in
docs/,specs/, or.scratch/directories (Step 2.3). - Boundary markers: The prompt instructions for the sub-agents in Step 4 do not include requirements for delimiters (such as XML tags or triple backticks) or specific instructions to ignore any embedded commands within the fetched spec or diff content.
- Capability inventory: The skill executes shell commands (
git) and delegates analysis to sub-agents. - Sanitization: There is no evidence of sanitization or filtering of the external data before it is interpolated into the prompts for the sub-agents.
- [COMMAND_EXECUTION]: The skill uses a user-supplied 'fixed point' (e.g., a branch name or commit hash) to construct and execute several git commands.
- Specifically, it executes
git diff <fixed-point>...HEAD,git log <fixed-point>..HEAD, andgit rev-parse <fixed-point>. If the user input is not properly sanitized by the underlying execution environment, it could potentially lead to command injection via shell metacharacters.
Audit Metadata