generating-audio
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides instructions for utilizing text-to-speech tools and includes best practices for sanitizing input text by stripping stage directions and markdown.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied scripts which are passed to the audio_generate tool.
- Ingestion points: User-provided script text in SKILL.md (Step 4).
- Boundary markers: Explicit instructions are provided to strip unspoken text, stage directions, and bracketed cues, and to use separate tool parameters for delivery style rather than text-based markers.
- Capability inventory: audio_generate, list_voices, and list_audio_models tools.
- Sanitization: Instructs the agent to filter out non-spoken content and markdown before generation.
- [SAFE]: No suspicious network activity, obfuscation, or unauthorized tool usage was detected. The skill utilizes standard tools for voice listing and audio generation within the platform's expected scope.
Audit Metadata