architecture
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of architectural guidelines and process instructions for an AI agent. It does not contain any executable code, network operations, or data exfiltration patterns. The instructions focus on improving code structure and security through principles like 'Authority' (server-side identity and ownership checks).
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to explore existing repository files during the 'Explore' phase to inform architecture decisions. While this process involves reading untrusted content from the codebase, the skill focuses on generating structural reports and code implementations within project boundaries. This behavior is inherent to the skill's primary purpose of code analysis and does not expose high-risk capabilities.
Audit Metadata