brooks-review
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill definition and the review guide consist strictly of markdown-based instructions for the agent to follow. There are no scripts (e.g., Python, JavaScript), shell commands, or external package dependencies included in the skill.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data in the form of code diffs and files provided by the user. This creates a potential surface for indirect prompt injection where malicious instructions could be embedded in code comments. However, the risk is mitigated as the skill does not grant the agent capabilities to perform sensitive actions like modifying the file system or making external network calls.
- Ingestion points: Pull request diffs, individual code files, and user-pasted code snippets as defined in the trigger conditions in SKILL.md.
- Boundary markers: The provided files do not specify explicit boundary markers or XML-style tagging instructions to isolate the untrusted code content.
- Capability inventory: The skill contains no tool calls, subprocess executions, or network-enabled functionality.
- Sanitization: The instructions do not include specific logic to sanitize or escape untrusted code content before the agent processes it.
Audit Metadata