speckit-review-comments
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local shell script at '.specify/scripts/bash/detect-changed-files.sh'. This is a design requirement for the spec-kit project structure used to detect changes for review.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from the contents of the files it reviews as it lacks boundary markers and sanitization.
- Ingestion points: The agent reads the contents of changed files located by the detection script.
- Boundary markers: Not present in the skill instructions to separate file data from instructions.
- Capability inventory: Execution of a local shell script via the agent's environment.
- Sanitization: No validation or escaping of the processed file content is specified.
Audit Metadata