The Agent Skills Directory

[SAFE]: The skill consists of instructional text for audio processing and utilizes internal workspace actions.
No malicious patterns, external downloads, or unauthorized system access were detected.
The author 'vidmoat' uses specific tool names like audio_inspect and setAudio which appear to be part of the intended environment.
[INDIRECT_PROMPT_INJECTION]: The skill describes a workflow that ingests audio metadata and descriptions to guide the agent's actions.
Ingestion points: Data is received via the audio_inspect action which returns measurements and audio-input descriptions.
Boundary markers: Not present, though instructions emphasize that measured timing and silence take precedence over model prose.
Capability inventory: The skill is limited to calling specific audio processing tools (setAudio, compress, eqLow, etc.) and does not have access to general shell execution or network requests.
Sanitization: Not applicable as the data is used to inform internal parameter settings for audio filters.

audio-cleanup