architecture
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill is an architectural design and review tool that uses structured protocols and adversarial validation to ensure safety and correctness. It emphasizes security best practices such as tenant isolation and secrets redaction.
- [COMMAND_EXECUTION]: The skill executes local utility scripts (e.g., node .claude/scripts/ai-signal-scan.cjs, python .claude/scripts/code_graph) to perform code analysis. These operations are limited to the local project environment and are part of the intended analytical functionality.
- [INDIRECT_PROMPT_INJECTION]: As a code review and audit tool, the skill processes untrusted project source code and documentation. It mitigates this risk through strict execution protocols, clear boundary markers in sub-agent prompts, and a mandatory validation gate ($why-review) to verify findings.
Audit Metadata