receiving-code-review
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes feedback from "External Reviewers," creating an attack surface where untrusted data enters the agent context.
- Ingestion points: External code review strings and GitHub pull request comment threads (mentioned in SKILL.md).
- Boundary markers: Absent. The instructions provide logical guidelines for evaluation but do not define technical delimiters to isolate external feedback from agent instructions.
- Capability inventory: The skill utilizes GitHub CLI tools (
gh api) and filesystem search (grep), and implies the ability to modify the codebase during the implementation phase. - Sanitization: Absent. The skill relies on the agent's behavioral evaluation ("verify before implementing") rather than technical sanitization or input filtering.
Audit Metadata