The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses ffmpeg for audio analysis and processing. In scripts/build_mix.py, the subprocess.run function is used to invoke ffmpeg for MP3 decoding. This execution is performed using a list of arguments and hardcoded filenames from internal tables, preventing command injection.
[EXTERNAL_DOWNLOADS]: The pipeline incorporates external APIs for transcription (OpenAI Whisper) and sound generation (ElevenLabs). These interactions are integral to the skill's primary purpose and do not involve downloading or executing arbitrary remote code.
[DATA_EXPOSURE]: Secret management (API keys) follows standard best practices by sourcing credentials from environment variables or .env files. No evidence of unauthorized file access or exfiltration of sensitive data was found.
[DYNAMIC_EXECUTION]: While the script performs complex audio manipulation using numpy and scipy, it does not utilize dangerous dynamic execution patterns such as eval(), exec(), or runtime code generation from untrusted sources.

ai-audio-sound-design