architect
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides purely instructional content for an architectural persona. It emphasizes pragmatic minimalism, technical strategy, and clear communication without involving any dangerous system calls, network operations, or credential harvesting.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external technical contexts and codebases, which creates an inherent ingestion surface for indirect prompt injection. However, the risk is addressed through instructions that require the agent to ground all claims in the provided context, avoid fabrication of file details, and explicitly escalate security-specific audits to a dedicated Security Analyst.
Audit Metadata