← 返回 Skills 市场
1598
总下载
1
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install llm-supervisor-agent
功能描述
Automatically switches between cloud and local LLMs on rate limits, with user confirmation required for local code generation.
使用说明 (SKILL.md)
LLM Supervisor
Automatically switches OpenClaw between cloud and local Ollama models when rate limits occur.
Features
- Detects rate-limit / overload errors from Anthropic/OpenAI
- Switches to a local Ollama fallback model
- Requires explicit confirmation before local code generation
- Supports manual commands:
Commands
/llm status/llm switch cloud/llm switch local
Default Local Model
qwen2.5:7b
Safety
Local code generation requires the user to type:
CONFIRM LOCAL CODE
安全使用建议
This skill is internally consistent with its stated purpose. Before installing: 1) Note it will set the agent LLM profile to your cloud provider ('anthropic:default') or to a local Ollama endpoint at http://127.0.0.1:11434 — ensure you run Ollama at that address or change the config. 2) It is enabled by default (skill.json enabledByDefault: true), so it may start affecting agents immediately; if you prefer control, disable it until tested. 3) It can notify all users and block code tasks in local mode until someone types the configured confirmation phrase, so review or change confirmationPhrase/localModel/cooldownMinutes if needed. 4) It does not request credentials or call external servers beyond a local Ollama URL; still review source if you require additional assurance. If you want extra caution, test in a non-production workspace first.
功能分析
Type: OpenClaw Skill
Name:
Developer:
Version:
Description: OpenClaw Agent Skill
Suspicious High-Entropy/Eval files: 12
The OpenClaw LLM Supervisor skill is designed to automatically switch between cloud and local LLM models (Ollama) when cloud rate limits are encountered. It includes a safety mechanism requiring explicit user confirmation (`CONFIRM LOCAL CODE`) before executing code-related tasks with a local LLM. The code primarily interacts with the OpenClaw SDK for state management, agent LLM profile configuration, and user notifications. There is no evidence of data exfiltration, malicious execution, persistence mechanisms, or prompt injection attempts in the code or documentation. The permissions requested (`llm`, `agents`, `notifications`, `state`) are appropriate for its stated functionality, and the local Ollama `baseUrl` is correctly set to `http://127.0.0.1:11434` in `hooks/onAgentStart.ts`.
能力评估
Purpose & Capability
The skill's name/description match the implementation: hooks detect LLM errors, switch a global mode state, set agent LLM profiles (cloud vs local), and expose /llm commands. Declared permissions (llm, agents, notifications, state) align with these actions.
Instruction Scope
Runtime behavior is narrowly scoped: it only inspects LLM error messages, agent start events, and task intents; it blocks code-related tasks in local mode until the user supplies the explicit confirmation phrase. It does not read arbitrary files, environment variables, or other agents' credentials.
Install Mechanism
No install spec or external downloads are present; the skill is packaged with its code (compiled dist files). There are no external URL downloads or extract steps. This is low-risk from an installer perspective.
Credentials
The skill requests no environment variables or external credentials. It does configure a local Ollama baseUrl (http://127.0.0.1:11434) when switching to local mode, which is appropriate for its purpose.
Persistence & Privilege
skill.json sets enabledByDefault: true, so the skill will be enabled by default; however 'always' is false and it does not request system-wide privileges beyond managing agents/llm profiles and state. Because it can change agent LLM profiles and notify all users, enable-by-default means it will affect agent behavior immediately unless you disable it.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install llm-supervisor-agent - 安装完成后，直接呼叫该 Skill 的名称或使用
/llm-supervisor-agent触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v0.2.0
Cloud/local LLM switching with rate limit detection and safety confirmation
元数据
常见问题
LLM Supervisor 是什么？
Automatically switches between cloud and local LLMs on rate limits, with user confirmation required for local code generation. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 1598 次。
如何安装 LLM Supervisor？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install llm-supervisor-agent」即可一键安装，无需额外配置。
LLM Supervisor 是免费的吗？
是的，LLM Supervisor 完全免费（开源免费），可自由下载、安装和使用。
LLM Supervisor 支持哪些平台？
LLM Supervisor 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 LLM Supervisor？
由 dhardie（@dhardie）开发并维护，当前版本 v0.2.0。
推荐 Skills