updating-knowledge
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection] (SAFE): The skill is designed to ingest untrusted web content. While this creates an indirect injection surface, it is the primary intended function and the instructions incorporate cross-validation steps to mitigate risks. * Ingestion points: Web content retrieved during the Research Workflow (SKILL.md). * Boundary markers: None specified in the instructions to separate fetched content from instructions. * Capability inventory: Includes web access and internal context tools (GitHub, Drive). * Sanitization: Relies on the methodology's cross-validation requirement of 3-5 diverse sources to verify information.
Audit Metadata