press-release-writing
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references external resources for installation and research, specifically pointing to the 'belt' CLI and documentation hosted on GitHub (inference-sh/skills). It also suggests expanding agent capabilities by adding related skills from the same ecosystem. These references are transparent and directly support the skill's stated purpose.
- [SAFE]: Analysis of the instructions and provided examples confirms that the skill focuses on journalistic standards and fact-checking. It does not contain prompt injection attempts, credential harvesting, or unauthorized network operations. Use of external search tools (Tavily, Exa) is explicitly defined for the purpose of verifying claims in the press releases.
Audit Metadata