roast-ja
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified. The skill is composed of purely instructional content.
- [PROMPT_INJECTION]: The instructions include persona-defining constraints, such as telling the AI not to act as a standard assistant and to avoid providing direct answers. These instructions are functional for the roasting persona and do not attempt to bypass safety filters or ignore system-level security constraints.
- [EXTERNAL_DOWNLOADS]: The skill includes a link to an external GitHub repository (tanweai/pua) as a recommendation for users. This is a static reference in the documentation and does not involve automated downloads or runtime execution of external code.
Audit Metadata