headlines
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified across any of the 11 threat categories. The skill is a template for natural language transformation (generating headlines from text) and adheres to safe practices.
- [INDIRECT_PROMPT_INJECTION]: Although the skill processes untrusted data (briefs and drafts), the instructions define a highly structured four-phase workflow that limits the surface for indirect injection. The availability of tools like Bash is standard for the agent's environment and is not exploited by the skill instructions.
Audit Metadata