Openai Whisper Andy27725

Local speech-to-text with the Whisper CLI (no API key).

Security Scan

The skill is internally consistent with a local Whisper CLI wrapper: it requires the whisper binary and suggests installing a brew formula; no credentials or unrelated accesses are requested—verify the brew formula source before installing.

Purpose & Capability

The name/description (local speech-to-text with Whisper CLI) matches the declared requirement (whisper binary) and the included install spec (brew formula that provides a whisper binary). There are no unrelated credentials, binaries, or config paths requested.

Instruction Scope

SKILL.md only instructs the agent to run the whisper CLI against local audio files and notes that models download to ~/.cache/whisper. It does not request reading unrelated files or environment variables. Note: model downloads are network activity and will write to disk (cache directory).

Install Mechanism

Install spec uses a brew formula (openai-whisper) which is a low-risk mechanism in general. However the formula's tap/source is not specified here; a third-party tap could install arbitrary code. Recommend verifying the formula's origin and audit the formula before installing.

Credentials

The skill requests no environment variables, credentials, or config paths — appropriate for a local CLI-based transcription tool.

Persistence & Privilege

always is false and the skill does not request permanent presence or elevated privileges. It will cause model files to be cached on first run but does not modify other skills or global settings.

Assessment

This skill appears to do what it claims: run the local Whisper CLI to transcribe audio. Before installing: (1) verify the Homebrew formula 'openai-whisper' comes from a trusted tap (inspect the formula and its source repository); (2) be aware that model files will be downloaded to ~/.cache/whisper (network activity and disk usage); (3) confirm the installed 'whisper' binary is the expected implementation (run whisper --version or check upstream repo); (4) note a minor documentation inconsistency: the SKILL.md mentions a default model named 'turbo', which is not a standard Whisper model name — treat that as a documentation error, not a credential or access red flag. If you are unsure about the brew formula's provenance, install in a sandbox/VM or build the upstream project from source instead.

Whisper (CLI)

Use whisper to transcribe audio locally.

Quick start

• whisper /path/audio.mp3 --model medium --output_format txt --output_dir .
• whisper /path/audio.m4a --task translate --output_format srt

Notes

• Models download to ~/.cache/whisper on first run.
• --model defaults to turbo on this install.
• Use smaller models for speed, larger for accuracy.