authority-ad
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill's Review Mode allows the agent to analyze external, user-supplied media via the $video-analysis tool. This ingestion point creates an attack surface for indirect prompt injection, where malicious instructions embedded in the analyzed content could attempt to influence the agent's output. The evidence chain includes: (1) Ingestion points: User-supplied ad in Review Mode in SKILL.md; (2) Boundary markers: Instructions in SKILL.md and references/shot-script-rubric.md to base output on observable evidence; (3) Capability inventory: $video-analysis in SKILL.md, and production batch runners in references/workflow-handoff.md; (4) Sanitization: No explicit content sanitization or escaping is described.
- [SAFE]: The skill incorporates comprehensive integrity checks, explicitly prohibiting the invention of credentials, studies, or testimonials, and setting clear boundaries against imitating existing creators or using unauthorized brand marks.
Audit Metadata