llm-wiki-faq
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists exclusively of Markdown documentation and instructional prompts for guiding AI responses. There are no executable scripts or system commands included.
- [SAFE]: The instructions include proactive safety gates and data classification policies. Specifically, it explicitly warns the agent to never ingest secrets and provides clear routing instructions to specialized privacy and security skills (e.g.,
llm-wiki-privacy-redactor,llm-wiki-security-review) when sensitive data is mentioned. - [SAFE]: All external URLs point to well-known academic and technology services, including arXiv, LangChain, and Anthropic. No typosquatting or suspicious redirection was found.
- [SAFE]: While the skill ingests untrusted user questions (Category 8 surface), its capabilities are limited to generating text responses based on the provided reference files, presenting no significant risk of indirect prompt injection affecting system integrity.
Audit Metadata