requesting-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 is designed to analyze external data such as code diffs and requirement documents, which inherently creates a surface for indirect prompt injection. The review subagent relies on structural markdown headers rather than strict delimiters to distinguish between instructions and the content being reviewed.\n
- Ingestion points: The workflow processes output from
git diffand text provided through placeholders for task descriptions and requirements.\n - Boundary markers: The template uses clear section headers (e.g.,
## What Was Implemented,## What to Check) to organize the prompt, but it does not include explicit instructions for the subagent to ignore commands potentially embedded within the code under review.\n - Capability inventory: The reviewer subagent is tasked with analysis and textual feedback; it is not granted access to sensitive files or destructive system tools within this workflow.\n
- Sanitization: Ingested code and requirements are passed directly into the subagent prompt without pre-analysis or character escaping.
Audit Metadata