The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes ffmpeg via execFileSync to concatenate audio segments and manage formats. The use of an argument array instead of a raw shell string provides protection against command injection.
[EXTERNAL_DOWNLOADS]: The instructions specify the installation of standard Python libraries (voxcpm, qwen-tts, soundfile) and the retrieval of AI model weights from established research repositories. These are required for the skill's primary audio generation functionality.
[SAFE]: The skill follows security best practices for secret management, instructing users to use environment variables for API keys rather than hardcoding them. Additionally, the inclusion of a smoke test script that performs privacy checks to identify accidental credential exposure indicates a high standard of security awareness.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied scripts from lines.json to generate audio. While this represents a data ingestion surface, the risk is negligible as the data is used solely as input for text-to-speech engines without being interpreted as agent instructions.

create-audio