architect
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill describes a systematic architectural design workflow (Ground, Sketch, Agree, Implement, Scrap) that emphasizes grounding in existing systems and evaluating designs against specific 'red flags' like shallow modules or information leakage.
- [INDIRECT_PROMPT_INJECTION]: The workflow involves ingesting data from existing subsystems (Phase A) and user-provided tasks to generate design sketches. While this constitutes an input surface, the process is internal to the agent's reasoning loop and includes multiple validation steps (Phase B screening, Phase C human sign-off) that act as natural mitigations against adversarial instructions in the input data.
- [COMMAND_EXECUTION]: The skill orchestrates other agent tools and skills (such as
how,why,arena, andinterrogate). These are referenced as internal capabilities for information gathering and design synthesis rather than execution of arbitrary external shell commands.
Audit Metadata