perimeter
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were found. The skill is designed for architectural auditing and specifically instructs the agent not to mutate files or propose code patches. It uses standard file and symbol search capabilities to analyze local source code.
- [INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface as it analyzes external codebases. This is considered safe as the skill lacks high-risk capabilities and performs no write operations. 1. Ingestion points: Target directory or subsystem files via symbol and file search tools. 2. Boundary markers: Absent; the skill does not define specific delimiters for separating code content from instructions. 3. Capability inventory: Limited to file/symbol searching and spawning research subagents. The skill explicitly prohibits file mutation. 4. Sanitization: Absent; code content is processed directly for architectural patterns.
Audit Metadata