Whisper GPU Audio Transcriber
v1.0.3Convert audio to SRT subtitles using OpenAI Whisper with automatic GPU acceleration for Intel XPU / NVIDIA CUDA / AMD ROCm / Apple Metal. Ideal for content c...
⭐ 0· 87·0 current·0 all-time
byAllan.M@allanmeng
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (Whisper GPU transcription) match the requested artifacts: a Python script that loads a local Whisper model and transcribes audio. Declared binaries (python) and suggested pip packages are appropriate for the stated purpose.
Instruction Scope
SKILL.md instructs the agent to run scripts/transcribe.py with a given audio path. The script only reads the provided audio file, loads a local model, transcribes to SRT, and writes the SRT next to the audio — it does not read unrelated files, export environment variables, or call remote endpoints in code.
Install Mechanism
This is instruction-only in the registry (no formal install spec), but SKILL.md instructs pip installs (openai-whisper and a PyTorch wheel/index). Those are expected. Note: the first run will download model files (~1.5GB) from the Whisper model host — this network activity is expected for local model use and may be large; mirrors or manual downloads are suggested in the docs for regions with restricted access.
Credentials
The skill does not request environment variables, credentials, or config paths. The script only accesses the specified audio file and writes an SRT file; no sensitive credentials are required.
Persistence & Privilege
always is false and the skill does not request permanent agent presence or modify other skills/system-wide settings. It runs a local script on demand.
Assessment
This skill appears to do exactly what it claims: run a local Whisper model to convert audio to SRT. Before installing or running it, consider: (1) first run will download large model files (~1–2 GB) to ~/.cache/whisper (network and disk usage); (2) you must install a PyTorch build that matches your hardware (the SKILL.md points to the official PyTorch index); (3) run the script in a controlled environment (virtualenv/container) if you are cautious about running third‑party Python code; (4) the code does not request credentials or exfiltrate data, but the model download comes from the internet — if you prefer, manually obtain model files from trusted mirrors and place them in ~/.cache/whisper. Overall the skill is internally consistent and coherent with its stated purpose.Like a lobster shell, security has layers — review code before you run it.
audiovk976wfkj0zhyb1fhmrrs32yh6h83hjwxcudavk976wfkj0zhyb1fhmrrs32yh6h83hjwxgpuvk976wfkj0zhyb1fhmrrs32yh6h83hjwxlatestvk97ee95yzx0e7m50v6dekvxc9s83j4frmetalvk976wfkj0zhyb1fhmrrs32yh6h83hjwxrocmvk976wfkj0zhyb1fhmrrs32yh6h83hjwxsubtitlevk976wfkj0zhyb1fhmrrs32yh6h83hjwxtranscriptionvk976wfkj0zhyb1fhmrrs32yh6h83hjwxwhispervk976wfkj0zhyb1fhmrrs32yh6h83hjwxxpuvk976wfkj0zhyb1fhmrrs32yh6h83hjwx
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🎙️ Clawdis
Binspython