pr-review-workflow
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process pull request (PR) data, which is an external, untrusted input source.
- Ingestion points: The agent is instructed to read the PR title, description, and the full code diff in Phase 0 (Pre-Review Setup) and Phase 2 (Deep Review).
- Boundary markers: The instructions do not specify explicit delimiters or "ignore instructions" prompts to isolate the PR content from the agent's logic, though it encourages the use of specialized security skills like
security-patterns-code-review. - Capability inventory: The skill orchestrates analysis tasks that involve reading files and potentially writing review comments. It does not explicitly grant the agent network or arbitrary command execution capabilities beyond standard PR review tools.
- Sanitization: No specific sanitization or filtering of the input PR text is described, relying instead on the agent's internal logic and safety guardrails during the review process.
Audit Metadata