llm-wiki-paf-adoption
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is primarily a reasoning and instructional guide for mapping LLM-Wiki to the Product Architecture Framework. It does not perform sensitive file access, execute arbitrary commands, or establish persistence. All instructions focus on logic, documentation, and user guidance.
- [PROMPT_INJECTION]: The skill includes instructions to process external data sources which represents a potential surface for indirect prompt injection. However, this is managed through human review states (draft vs. verified) and explicit provenance requirements. 1. Ingestion points: External PAF source pages and user-provided citations. 2. Boundary markers: The skill emphasizes separating source-backed facts from human synthesis and generated summaries. 3. Capability inventory: The skill mentions triage and ingestion workflows but limits its own scope to reasoning. 4. Sanitization: Human review is the primary validation mechanism defined.
Audit Metadata