The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads a pre-trained RNNoise model from a public repository (https://github.com/GregorR/rnnoise-models/raw/master/somnolent-hogwash-2018-09-01/sh.rnnn) using curl. This model is required for Tier 3 denoising functionality.\n- [COMMAND_EXECUTION]: The skill executes a local Python script (skills/shared/scripts/audio_ops.py) to interface with ffmpeg. This is the primary mechanism for the skill's functionality, used for probing file information and applying audio filters.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided audio/video files and parameters (tier, mix). This creates a potential attack surface for indirect prompt injection or command injection via malformed filenames or metadata.\n
Ingestion points: input_file path and user-specified parameters in SKILL.md.\n
Boundary markers: None specified.\n
Capability inventory: Execution of local scripts and subprocess calls to ffmpeg.\n
Sanitization: Not explicitly described; the skill relies on a wrapper script for command construction.

audio-denoise