code-review
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses user-supplied strings directly in shell commands, such as
git diff <fixed-point>...HEADandgit rev-parse <fixed-point>. This creates a potential surface for command or argument injection if the platform does not strictly validate that the input is a valid git reference before execution. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from the repository, including commit messages, file contents in diffs, and external specification files.
- Ingestion points: Repository data entered into the context via
git diff,git logoutput, and local documentation files (docs/,specs/). - Boundary markers: The instructions lack explicit delimiters (e.g., XML tags or triple quotes) or 'ignore' instructions to isolate the data being reviewed from the agent's logic.
- Capability inventory: The agent has the ability to execute shell commands (
git) and read files from the project directory. - Sanitization: No validation or sanitization of the analyzed code or metadata is performed before it is passed to the parallel sub-agents for review.
Audit Metadata