ClawHub

OpenClaw Voice Control

v0.1.9

Local macOS voice-control integration for OpenClaw. Use when setting up, deploying, troubleshooting, or operating wakeword-triggered voice access to a local...

⭐ 0· 249·0 current·0 all-time
by@carrotyuan
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Pending
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name, description, required binaries (python3, git, launchctl), required env vars (OPENCLAW_BASE_URL, OPENCLAW_TOKEN, SENSEVOICE_MODEL_PATH, SENSEVOICE_VAD_MODEL_PATH) and launchd usage align with a local macOS voice-control integration for OpenClaw.
ℹ
Instruction Scope
SKILL.md directs the agent to clone the GitHub repo into the skill workspace, create/activate a venv, pip install -e ., download local models, run the voice service and an overlay, and optionally deploy a launchd auto-start. These steps are within scope but involve running repository code and network downloads; the skill does explicitly require user confirmation before system-changing steps.
✓
Install Mechanism
There is no automated install spec in the registry (instruction-only). The recommended flow uses git fetch and pip install -e . from the upstream GitHub repo and model downloads (modelscope/funasr). This is expected for this kind of local integration, but pip install will execute the repo's installation code so the repo contents should be reviewed prior to installation.
✓
Credentials
Requested environment variables are appropriate for the stated purpose: a base URL and token to contact the local OpenClaw agent and local model path variables for ASR/VAD. No unrelated credentials or broad cloud creds are requested. Note that OPENCLAW_TOKEN is a primary credential that grants whatever access the OpenClaw service allows, so use least-privilege tokens.
ℹ
Persistence & Privilege
The skill does not force permanent inclusion (always:false). It explicitly offers optional background resident behavior and launchd auto-start via a deploy script; this is reasonable for a persistent voice entrypoint but increases attack surface and should only be enabled after inspecting deploy_macos.sh and confirming intent.
Assessment
This skill appears to do what it says, but it will fetch and install code from the linked GitHub repository and download local ML models. Before installing: (1) review the repository (especially setup/pyproject, setup scripts, and scripts/deploy_macos.sh) to confirm there is no unexpected behavior, (2) do not supply a highly-privileged OPENCLAW_TOKEN — create a least-privilege token for voice access, (3) confirm any steps that modify system state (launchd auto-start), and (4) prefer running the initial validation in an isolated user account or VM if you want to limit risk. If you are not comfortable reviewing the repo, do not run pip install -e . or enable auto-start.

Like a lobster shell, security has layers — review code before you run it.

latestvk97dapat0dck2j8yfkfhef2eas83jfwd

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Runtime requirements

OSmacOS
Binspython3, git, launchctl
EnvOPENCLAW_BASE_URL, OPENCLAW_TOKEN, SENSEVOICE_MODEL_PATH, SENSEVOICE_VAD_MODEL_PATH
Primary envOPENCLAW_TOKEN

Comments