Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Faster Whisper Gpu
v0.1.0High-performance local speech-to-text transcription using Faster Whisper with NVIDIA GPU acceleration. Transcribe audio files locally without sending data to...
⭐ 0· 465·0 current·0 all-time
byFelipe Oliveira@felipeoff
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description, SKILL.md, requirements.txt and transcribe.py all align: the skill implements local Faster-Whisper transcription with optional GPU support. Required binary is only python3; listed Python packages match the stated capability.
Instruction Scope
SKILL.md and transcribe.py limit actions to local transcription, writing outputs to stdout or local files. The only external activity is the expected download of model weights from the Hugging Face hub (~ ~/.cache/huggingface/hub) on first use; this is documented in SKILL.md. No instructions read unrelated system credentials or post audio to third-party endpoints.
Install Mechanism
There is no registry install spec; the skill is instruction-only and instructs users to pip install faster-whisper and torch. Using pip is expected but may pull large binary wheels (torch) and GPU-specific builds. Model weights are downloaded at runtime from Hugging Face — an expected but noteworthy network activity and storage use.
Credentials
The skill requests no environment variables or credentials. SKILL.md mentions HF_HOME as an optional way to change the Hugging Face cache directory (documented as a troubleshooting tip) — reasonable and proportional.
Persistence & Privilege
The skill is not always: true and does not request persistent system privileges. It does not modify other skills or agent-wide configs. It writes outputs and caches model files to the user's home cache directory (standard behavior).
Assessment
This skill appears to be what it says: a local Faster-Whisper transcription tool. Before installing, consider the following: (1) pip installing torch can be large and may require specific GPU/CUDA builds — prefer installing inside a virtualenv or conda environment. (2) Model weights are downloaded from the Hugging Face hub on first use (network access and several GBs of disk may be required for large models); if you require full offline operation, pre-download and place models in the HF cache or set HF_HOME. (3) Review the GitHub homepage/repo and the transcribe.py file if you need additional assurance about code behavior and licensing. (4) The skill does not request credentials or attempt to send your audio elsewhere, but network activity for model downloads is expected. If those behaviors are acceptable, the skill is coherent and reasonable to use.Like a lobster shell, security has layers — review code before you run it.
latestvk97cj6xwf2rx7gvr17t8wpy7s581k7qs
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🎙️ Clawdis
Binspython3