Shelly

Control and automate Shelly devices with local RPC workflows, secure access modes, cloud API coordination, and safe multi-device execution.

This skill appears coherent and focused on Shelly device control. Before installing: (1) only provide a SHELLY_CLOUD_TOKEN scoped to the minimum permissions needed and rotate it when possible; (2) keep the ~/shelly/ folder permissions as recommended (700/600) and avoid storing raw tokens there; (3) confirm curl and jq are trusted versions on your system; (4) be aware the agent can be invoked normally (autonomously) — enable write/apply modes only when you trust the workflow and have canary safeguards enabled; (5) if you use a cloud token, consider creating a token limited to the devices or actions you need rather than a broad account token.

Type: OpenClaw Skill Name: shelly Version: 1.0.0 The Shelly skill bundle is a well-structured tool for managing Shelly IoT devices via local RPC, WebSocket, and Cloud APIs. It includes robust safety protocols such as read-before-write verification, canary rollout strategies, and explicit user confirmation for high-impact actions (e.g., toggling relays). The skill manages its state locally in `~/shelly/` with appropriate file permissions and explicitly directs the agent to use environment variables for sensitive tokens rather than storing them in plaintext, showing a clear focus on security and operational reliability.