using-experiment-formalisation
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of markdown documentation and JSON data files providing guidance on experiment formalisation. No executable code or malicious scripts were found.
- [EXTERNAL_DOWNLOADS]: The skill references several external resources, including SourceForge, Hugging Face, and GitHub repositories for the EXPO and SUMO ontologies. These are well-known or trusted services, and the references are for informational purposes or to fetch legitimate ontology artifacts.
- [PROMPT_INJECTION]: No prompt injection patterns were detected. The instructions provided are meta-instructions for an agent to follow to ensure accuracy and prevent common pitfalls like hallucinating vocabulary (e.g., Law 1: Verified vocabulary only). These are benign and intended for task-specific guidance.
- [INDIRECT_PROMPT_INJECTION]: The skill specifically addresses and mitigates the risk of hallucinated vocabulary and improper data modeling by requiring primary-source verification of terminology, which improves the accuracy and safety of the agent's outputs.
Audit Metadata