Security audit

ZhiHand

Security checks for vulnerabilities and agentic risk

Overview

ZhiHand appears to do what it claims—pair a mobile device with OpenClaw for screen reading, chat relay, and phone control—but it requires a sensitive OpenClaw gateway token and access to phone-screen/control data.

This looks internally consistent for a phone-control bridge. Before installing, make sure you actually want a paired ZhiHand mobile app to send prompts/media into your OpenClaw agent and allow the agent to read screenshots and send phone actions. Treat the OpenClaw gateway token like a password, prefer enabling the ZhiHand tools only for a dedicated mobile agent, and verify that gatewayResponsesEndpoint stays local unless you intentionally self-host it. Confidence is medium because some source files in the provided artifact were truncated or omitted from review.

Static analysis

Detected: suspicious.exposed_secret_literal, suspicious.install_untrusted_source

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: src/plugin.ts:815
Evidence: authToken: [REDACTED](api),

Install source points to URL shortener or raw IP.

Warn

Code: suspicious.install_untrusted_source
Location: openclaw.plugin.json:36
Evidence: "default": "http://127.0.0.1:18789/v1/responses"