← Back to Skills Marketplace
openclaw-connect子节点专用skill
by
Scarlett_AI
· GitHub ↗
· v1.0.10
· MIT-0
244
Downloads
0
Stars
1
Active Installs
11
Versions
Install in OpenClaw
/install openclaw-connect-node
Description
OpenClaw Connect Enterprise 子节点。将当前 OpenClaw 实例注册为远程 Hub 的子节点，通过 appId/key/token 三重验证绑定。提供本地管理界面（任务查看、记忆管理、系统监控）。子节点只能看自己的数据，所有被主节点控制的数据不可操作。当用户说"连接主节点"、"注册子...
Usage Guidance
This package contains real server/frontend code and an installer but the registry metadata incorrectly says it's instruction-only and lists no env vars — do not trust metadata alone. Before installing: (1) inspect the full code (server/src/index.ts and install-node.sh) yourself or run in an isolated VM/container, (2) be aware it requires HUB_URL, APP_ID/APP_KEY/APP_TOKEN and will persist them to ~/.openclaw-node/node.json, (3) the node will scan ~/.openclaw for workspace files (SOUL.md, IDENTITY.md, skills) and may read ~/.openclaw/config.yaml — if those contain secrets you do not want exposed to the Hub, do not install on that host, (4) avoid piping remote scripts (curl | bash) and verify the git repo/source before cloning, and (5) if you still want to run it, use a dedicated host or container with minimal privileges and network access limited to the intended Hub.
Capability Assessment
Purpose & Capability
The description (register node to Hub, local UI, task execution) aligns with the code. However the registry metadata claims no required env vars and 'instruction-only', yet the bundle contains server/frontend code and requires HUB_URL / APP_ID / APP_KEY / APP_TOKEN (both in SKILL.md and server code). That metadata / manifest mismatch is an incoherence — the skill is not instruction-only and does require sensitive credentials.
Instruction Scope
Runtime instructions and server code go beyond simply exposing HTTP endpoints: the node persists credentials to ~/.openclaw-node/node.json, auto-registers local agents by reading ~/.openclaw/workspace/SOUL.md and IDENTITY.md and listing ~/.openclaw/workspace/skills, and attempts to inspect ~/.openclaw/config.yaml (via execSync('cat ...')). These file reads are related to agent registration but may access user data the installer/operator did not expect. The SKILL.md also includes examples that pipe a remote install script (curl | bash) — a risky install pattern if used.
Install Mechanism
The registry lists no install spec but the package contains install-node.sh which performs cloning (default GIT_REPO points to an external code host), installs Node.js via nodesource curl scripts, runs npm installs and builds, and may register systemd services. There is no single signed release URL; cloning and remote-install examples are present. This is not necessarily malicious but is higher-risk than a pure instruction-only skill.
Credentials
The skill bundle and SKILL.md require HUB_URL, APP_ID/APP_KEY/APP_TOKEN, but the registry metadata declares no required environment variables or primary credential. That mismatch is significant: the skill needs credentials (and will persist them to disk) but did not declare them in metadata. Additionally the code reads other user-local config (~/ .openclaw/*) which could surface channel configuration — this is proportionate to auto-registering local agents but worth noting as a privacy/credential exposure surface.
Persistence & Privilege
The node persists credentials to ~/.openclaw-node/node.json and the installer expects to create files under /opt/openclaw-node and register a systemd service (requires root). The skill does not set always:true. It does not appear to modify other skills' configs, but it will run as a long-lived service and write persistent credentials and data on the host — appropriate for a node service but a material persistence risk if deployed on a machine with other sensitive workloads.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install openclaw-connect-node - After installation, invoke the skill by name or use
/openclaw-connect-node - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.10
Version 1.0.10
- No file changes detected in this release.
- Documentation, code, and functionality remain unchanged from the previous version.
v1.0.9
No file changes detected in this version.
- No updates or modifications were made to the skill files.
- Documentation, features, and functionality remain unchanged from the previous version.
v1.0.8
Version 1.0.8 of openclaw-connect-node
- No file changes detected in this version.
- Documentation and instructions remain the same as previous version.
v1.0.7
- 文档统一将示例中的 Hub 地址由公网 IP 格式改为 YOUR_HUB_IP，更加通用。
- 架构图与环境变量表中的 Hub 地址也统一为 YOUR_HUB_IP，保持一致性。
- 其他内容未更改。
v1.0.6
- Major refactor: Reorganized the project structure for clarity and modularity.
- Added separate frontend and server directories with clean source code structure.
- Updated installation and usage instructions to reflect new layout.
- Replaced old asset-based frontend build with a source-based frontend package under `frontend/`.
- Improved documentation for setup, configuration, and API usage.
- Provided install and uninstall scripts for easier management.
v1.0.5
1. No code or documentation changes detected in this version.
2. Functionality, installation methods, and usage instructions remain the same as the previous release.
v1.0.4
Version 1.0.4
- Updated scripts/src/index.ts (exact code changes not specified).
- No user-facing or documentation changes detected.
v1.0.3
Version 1.0.3 of openclaw-connect-node
- No file changes detected in this release.
- No updates or modifications made to features, code, or documentation.
v1.0.2
openclaw-connect-node 1.0.2
- No changes detected in this version; all files and documentation remain unchanged.
v1.0.1
Version 1.0.1
- No file changes detected in this release.
- No updates or modifications to functionality or documentation.
v1.0.0
Initial release of openclaw-connect-node.
- Adds skill to register an OpenClaw instance as a secure sub-node for a remote Hub.
- Provides local management UI for tasks, memory, and monitoring with access restricted to the node’s own data.
- Includes quick install script, systemd integration for auto-start and resource limits, and command-line management tools.
- Implements heartbeat, periodic sync with the Hub, and robust reconnect logic.
- Offers clear permission boundaries: nodes cannot manage agents or other nodes, and all critical operations are reported to the Hub.
Metadata
Frequently Asked Questions
What is openclaw-connect子节点专用skill?
OpenClaw Connect Enterprise 子节点。将当前 OpenClaw 实例注册为远程 Hub 的子节点，通过 appId/key/token 三重验证绑定。提供本地管理界面（任务查看、记忆管理、系统监控）。子节点只能看自己的数据，所有被主节点控制的数据不可操作。当用户说"连接主节点"、"注册子... It is an AI Agent Skill for Claude Code / OpenClaw, with 244 downloads so far.
How do I install openclaw-connect子节点专用skill?
Run "/install openclaw-connect-node" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is openclaw-connect子节点专用skill free?
Yes, openclaw-connect子节点专用skill is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does openclaw-connect子节点专用skill support?
openclaw-connect子节点专用skill is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created openclaw-connect子节点专用skill?
It is built and maintained by Scarlett_AI (@ai-scarlett); the current version is v1.0.10.
More Skills