fact-check-before-trust
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill contains only Markdown documentation and metadata. No scripts, binaries, or other executable files are provided, which minimizes the technical attack surface.- [SAFE]: The logic described in the skill is defensive in nature, focusing on fact-checking and claim verification. No patterns of prompt injection, data exfiltration, or persistence were detected.- [INDIRECT_PROMPT_INJECTION]: The skill protocol involves processing agent-generated output, which is a surface for indirect instructions. Ingestion points: Agent output containing claims (SKILL.md). Boundary markers: None specified in the extraction protocol. Capability inventory: None (no tools or scripts). Sanitization: None described for the extracted claims. The risk is assessed as safe given the lack of executable capabilities.
Audit Metadata