design
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines conceptual workflows and instructions for 'NKS Design,' a framework for modeling systems as graphs of 'phenomena' and 'kriyas.' The content is entirely instructional and operates within the context of defined agent tools.
- [SAFE]: The instructions focus on architectural principles (e.g., 'Complex systems grow,' 'Tensions are truthful') and multi-phase design processes (Backward chaining, Forward weaving, Risk analysis).
- [SAFE]: While Category 8 (Indirect Prompt Injection) surface area is technically present because the skill mentions 'intake' from external sources and brainstorming suites, it does so to define data modeling boundaries rather than execution paths. No dangerous interpolation of untrusted data into shell commands or code execution was found.
- [SAFE]: The 'Risk Analysis' phase (Phase 3) actually encourages defensive thinking, instructing the agent to evaluate scale mismatches, adversaries, and state history.
- [SAFE]: No obfuscation, hardcoded credentials, malicious remote downloads, or persistence mechanisms were detected. The technical terms used (e.g., 'anagata', 'vimarsha', 'triputi') are domain-specific terminology for the NKS framework and are used for data categorization.
Audit Metadata