llm-wiki-human-first-design
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional and consists of Markdown guidelines and local documentation references. It does not include any executable scripts, binary files, or tools that interact with the host system or network.
- [SAFE]: The documentation provides robust security recommendations, including guidance on secret management (e.g., advising against ingesting API keys), data classification (Public, Internal, Confidential), and the implementation of human-in-the-loop review processes for AI-generated content.
- [SAFE]: The skill includes explicit defensive strategies against common AI risks like prompt injection and context poisoning by recommending immutable raw sources, claim-level provenance (linking back to original evidence), and protected human synthesis sections that an agent should not overwrite.
Audit Metadata