deprecated ignore
Analysis
The skill appears to do what it claims—bridge voice text through Hotbutter and run the local OpenClaw agent—but users should be comfortable with agent prompts and replies passing through a hosted relay.
Findings (3)
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.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
const args = [
'agent',
'--session-id', sessionId,
'-m', text,
];
...
execFile('openclaw', args, { timeout: 120_000 },The skill turns incoming text into a local OpenClaw agent invocation. This matches the documented voice-bridge purpose, but users should understand that paired relay messages can cause local agent turns.
"dependencies": {
"ws": "^8.18.0"
}The package depends on the WebSocket library using a semver range rather than an exact pinned version. This is expected for a Node WebSocket client, but dependency provenance still matters.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
Voice transcripts and agent responses are transmitted through the hosted relay server at `wss://hotbutter.ai`.
The skill intentionally sends transcribed speech and agent response text through a third-party hosted relay. This is clearly disclosed and central to the hosted voice feature.
