Chen Openai Whisper
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill coherently installs and runs the local Whisper transcription CLI, with only expected notes about external installation/model downloads and minor provenance ambiguity.
This appears safe for its purpose if you are comfortable installing the Whisper CLI and downloading models locally. Before installing, confirm the package/source is the one you expect, and only transcribe audio files you are comfortable processing on this machine.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing or first using the skill may download software and model files onto the local machine.
The skill depends on an external Homebrew package and first-run model downloads. This is disclosed and necessary for local Whisper transcription, but it is still a supply-chain dependency users should trust before installing.
"install":[{"id":"brew","kind":"brew","formula":"openai-whisper","bins":["whisper"]}] ... Models download to `~/.cache/whisper` on first run.Install only from a trusted Homebrew setup, expect the model cache download, and verify the package source if supply-chain integrity is important.
The skill's package identity is not perfectly consistent across the provided metadata.
The embedded metadata identity differs from the registry metadata shown for this evaluation, which lists a different owner and slug. This creates minor provenance ambiguity, though no unsafe behavior is shown.
"ownerId": "kn70pywhg0fyz996kpa8xj89s57yhv26", "slug": "openai-whisper"
Confirm that the listed publisher and package are the ones you intend to install.
The agent may run Whisper locally and create transcript files in the chosen output directory.
The skill instructs the agent to run a local CLI command on user-provided audio and write output files. This command execution is central to the stated transcription purpose and is not hidden or automatic.
`whisper /path/audio.mp3 --model medium --output_format txt --output_dir .`
Use the skill only on audio files you intend to transcribe, and specify an output directory where transcript files are expected.