architecture
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- NO_CODE (SAFE): The skill contains no executable scripts (Python, JavaScript, Shell, etc.). All functionality is derived from natural language instructions and Markdown templates.
- DATA_EXFILTRATION (SAFE): There are no commands or functions that perform network requests (curl, wget, etc.) or access sensitive file paths (~/.ssh, ~/.aws). All data processing is confined to the agent's internal reasoning and text output.
- PROMPT_INJECTION (SAFE): The instructions are focused on professional architectural design and do not contain patterns typical of prompt injection or attempts to bypass agent safety guidelines.
- INDIRECT_PROMPT_INJECTION (LOW): While the skill processes user-provided architectural requirements (untrusted data), it lacks any 'write' or 'execute' capabilities that could be exploited via injection. It only produces text-based documentation (ADRs, design specs).
Audit Metadata