code-review-guide
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data (pull requests and code diffs), which presents a surface for indirect prompt injection attacks where malicious instructions hidden in the code could attempt to influence the agent's output.
- Ingestion points: Untrusted data enters the agent context through the 'The change' (diff/PR) and 'Context' inputs defined in the 'Required Inputs' section of
SKILL.md. - Boundary markers: The instructions lack explicit boundary markers or delimiters to isolate user-provided code from the agent's operating instructions.
- Capability inventory: The skill is purely instructional for text generation; it does not request or utilize any tools for file system access, network operations, or shell command execution.
- Sanitization: There are no instructions for the agent to sanitize, escape, or filter the content of the diffs it reviews.
Audit Metadata