Qwen3 Audio
ReviewAudited by ClawScan on May 10, 2026.
Overview
The skill matches its audio TTS/STT purpose, but it can automatically install an unpinned prerelease Python dependency at runtime and stores reusable voice profiles, so it should be reviewed before use.
Before installing, confirm you are comfortable with the skill installing `mlx-audio` from the package ecosystem at runtime, preferably pin and preinstall the dependency yourself, and treat any voice cloning or saved voice profiles as sensitive data that should only be used with permission.
Findings (2)
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.
Using the skill may modify the local Python environment and execute newly downloaded package code before the user has reviewed exactly what version will run.
If the dependency is missing, the skill invokes a shell command to install a mutable prerelease dependency at runtime rather than relying on a pinned, reviewed install step.
except ImportError:
print("✗ mlx-audio 未安装，正在安装...", file=sys.stderr)
os.system("uv add mlx-audio --prerelease=allow")
import mlx_audioPin dependencies with a lockfile, declare `uv` and installation requirements in metadata/install specs, and remove automatic runtime installation or require explicit user approval before installing.
A created voice profile may leave a local copy of someone’s voice sample and transcript that can be reused later.
Reusable voice profiles persist reference audio and transcript/style text for later use by the skill.
Voices are stored in the `voices/` directory at the skill root level. Each voice has its own folder containing: - `ref_audio.wav` - `ref_text.txt` - `ref_instruct.txt`
Only create voice profiles with consent, avoid storing sensitive recordings unnecessarily, and delete the relevant `voices/` folder when no longer needed.