ds-review-loop
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard version control commands (
git status,git fetch,git merge-base,git diff) to identify changes and scope the review process. - [INDIRECT_PROMPT_INJECTION]: The skill ingests feedback from an external reviewer (another model or subagent), which constitutes an indirect prompt injection attack surface.
- Ingestion points: The agent parses the reviewer's response for specific findings, severities, and verdicts.
- Boundary markers: The instructions define a structured reply format for the reviewer but do not implement strict delimiters to isolate the reviewer's output from the agent's internal instructions.
- Capability inventory: The agent is authorized to modify local files to apply suggested fixes and execute shell commands to run verification checks.
- Sanitization: Risk is mitigated by instructions requiring the agent to independently judge each finding, verify fixes with evidence, and refer persistent disputes to the user.
Audit Metadata