Eagle Claw

分布式 AI 工作节点技能 - 连接星联(Skynet)调度系统，自动接单与执行任务

This skill asks the agent to connect to an external WebSocket and automatically accept and execute tasks from that network, and it will generate or accept an Ed25519 private key. Before installing, consider: (1) Do you trust the Skynet endpoint (SKYNET_WS_URL)? A malicious endpoint can send tasks that cause data leaks or unsafe actions. (2) Where will the generated PRIVATE_KEY be stored and who can read it? The SKILL.md doesn't say — ask for the storage location and protection measures. (3) The skill references tools but provides no code or install steps, so behavior is underspecified; request the implementation or a clear runtime spec. (4) If you proceed, run this skill in a restricted/sandboxed environment and avoid supplying a high-value PRIVATE_KEY until you verify the implementation. If you need stronger assurance, ask the publisher for source code, an install script, and documentation on how results and logs are transmitted and stored.

Type: OpenClaw Skill Name: skill-12 Version: 0.1.0 The eagle-claw skill (SKILL.md) is designed to transform the OpenClaw agent into a remote worker node for an external distributed network called 'Skynet' via a WebSocket connection (SKYNET_WS_URL). This architecture functions as a Command and Control (C2) mechanism, explicitly instructing the AI agent to 'automatically receive and execute tasks' from an external source. While the stated purpose is distributed AI collaboration, this setup introduces a high-risk remote code execution (RCE) vector, effectively allowing a third party to control the agent's actions and access the host environment.