codebase-context
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for information gathering and synthesis within an organization's internal environment.
- [PROMPT_INJECTION]: The skill handles untrusted data from code and documentation search results. It includes specific guidelines for the agent to 'be skeptical' and 'vet all content' based on relevance, currency, and authority, which serves as a mitigation against indirect prompt injection.
- [DATA_EXFILTRATION]: The skill accesses internal codebase data and employee information. This access is consistent with its purpose of architectural discovery. No network operations or data transmission to external domains were identified.
Audit Metadata