review-code
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data from various sources such as local source code, git diffs, and GitHub pull request descriptions.\n
- Ingestion points: The skill identifies targets for review including
diff,pr <number>, commit ranges, and entire codebases as specified inSKILL.md.\n - Boundary markers: The instructions do not provide explicit delimiters or instructions to treat the analyzed code as data rather than instructions, which is a common surface for indirect prompt injection.\n
- Capability inventory: The skill's capabilities include reading file system content, using the GitHub CLI to fetch PR data, invoking subagents for large reviews, and calling a secondary
second-opinionskill.\n - Sanitization: No specific content sanitization or validation routines are defined to mitigate potential instructions embedded within the codebase or PR metadata being reviewed.
Audit Metadata