Mijia

Control Xiaomi Mijia smart home devices. Use this skill when the user wants to control desk lamps, smart plugs, or other Mijia devices. Supports turning lights on/off, adjusting brightness, setting color temperature, switching modes, and more.

Before installing or running this skill, consider the following: - The skill will require you to set MIJIA_LAMP_DID and to login to your Xiaomi account (QR code). Confirm where the mijiaAPI library stores login tokens/credentials on disk (cookies, config files) and whether you are comfortable with that storage location. - The registry metadata does NOT list MIJIA_LAMP_DID as a required env var, and SKILL.md/README use inconsistent directory paths (~/.clawdbot vs ~/.claude vs /path/to). Ask the author to fix metadata and docs so requirements are explicit. - The skill depends on the third-party package 'mijiaapi' (pyproject.toml). Inspect that package's source or its PyPI/GitHub repo for any network endpoints, telemetry, or unexpected behaviors before running 'uv sync'. - Run the skill in an isolated environment (virtualenv or disposable VM) the first time to observe what files/tokens it creates and what network calls it makes. - If you only want the AI to issue commands and not perform autonomous installs/logins, avoid allowing autonomous invocation until metadata and behavior are clarified. If the author can (1) list MIJIA_LAMP_DID in the skill metadata, (2) fix the path inconsistencies, and (3) document where login tokens are stored (and confirm no unrelated credentials are accessed), this would reduce the outstanding concerns.

Type: OpenClaw Skill Name: mijia Version: 0.1.0 The skill is designed to control Xiaomi Mijia smart home devices, and all analyzed files align with this stated purpose. The `SKILL.md` provides clear instructions for the AI agent to execute commands via `scripts/lamp_cli.py`, which interacts with the `mijiaAPI` library. The script reads the `MIJIA_LAMP_DID` environment variable for device identification and calls `api.login()` for authentication, both of which are necessary and explicitly documented for smart home device control. There is no evidence of prompt injection with malicious intent, data exfiltration to unauthorized endpoints, malicious execution, persistence mechanisms, or obfuscation. The dependency on `mijiaapi` is declared in `pyproject.toml` and is central to the skill's functionality.