ai-code-reviewer
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted AI-generated code, which presents a surface for indirect prompt injection attacks.
- Ingestion points: The skill takes AI-generated code from external tools like Claude, ChatGPT, and Copilot as input.
- Boundary markers: The instructions provide explicit safety guidance, requiring execution only in a sandbox or throwaway checkout without credentials or production data.
- Capability inventory: The workflow involves the execution of tests and commands on the untrusted input.
- Sanitization: No specific automated sanitization is mentioned; the skill relies on the agent to follow sandboxing protocols and credential isolation during verification.
Audit Metadata