Automatically switches OpenClaw between cloud and local Ollama models when rate limits or overload errors occur.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install llm-supervisor-agent
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install llm-supervisor-agent using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is LLM Supervisor?
LLM Supervisor is a specialized productivity utility designed to ensure your AI development workflow never hits a standstill. By acting as an intelligent watchdog for Openclaw Skills, this tool monitors API responses from major cloud providers like Anthropic and OpenAI. When it detects rate-limit triggers or service overloads, it instantly transitions your agent to a local Ollama environment.
This skill provides a robust safety net for developers who rely on consistent availability. It bridges the gap between high-performance cloud intelligence and the reliability of local hardware, ensuring that Openclaw Skills remain functional even when external APIs are unavailable or exhausted. With a focus on both automation and safety, it is an essential component for high-frequency coding sessions.
LLM Supervisor Use Cases
- Preventing development downtime when reaching Anthropic or OpenAI API quotas.
- Maintaining coding progress during cloud provider outages or service maintenance.
- Manually switching to local models for offline development or cost-saving measures using Openclaw Skills.
- Testing code generation across different model architectures using local fallback instances.
How LLM Supervisor Works
- The supervisor monitors the communication stream between the agent and cloud LLM providers for 429 (Rate Limit) or 503 (Overload) errors.
- Upon detection of a service failure, the skill triggers an automatic failover to the local Ollama service.
- The system reconfigures the active model to the default local model (qwen2.5:7b) to maintain continuity.
- To ensure security, the agent pauses before any local code generation, requiring the user to type a specific confirmation string.
- Users can interact with the supervisor at any time via slash commands to check status or manually toggle between cloud and local modes.
LLM Supervisor Setup
To integrate this fallback mechanism into your Openclaw Skills setup, ensure Ollama is installed and the required local model is available:
# Install and pull the default local model
ollama pull qwen2.5:7b
# The skill will automatically detect the local Ollama instance on its default port
Once the model is pulled, the supervisor will begin monitoring cloud requests automatically.
LLM Supervisor Data Schema & Taxonomy
| Feature | Specification |
|---|---|
| Default Local Model | qwen2.5:7b |
| Detectable Errors | Rate-limit (429), Overload (503) |
| Safety Override | CONFIRM LOCAL CODE |
| CLI Interface | /llm [status |
| Cloud Compatibility | Anthropic, OpenAI |
LLM Supervisor Advanced Features
- Intelligent failover automation that detects cloud instability in real-time.
- Manual environment control via a dedicated slash command system within Openclaw Skills.
- Security-first local execution that requires explicit human confirmation for code generation.
- Support for custom local model configurations to replace the default qwen architecture.
SKILL.md
Loading
Related Openclaw Skills
Skill-Scan: Security Auditor for Agent Skills
A multi-layered security scanner designed to audit Openclaw Skills for malicious code, prompt injection, and behavioral threats before they are enabled.
Claude Chrome
A bridge that enables Claude Code to interact with the Chrome browser for advanced web automation and data retrieval.
Pndr
Pndr is a comprehensive personal productivity command center that integrates with AI agents via MCP to manage tasks, habits, and journals through natural language.
Agent Communication Skill
A high-performance WebSocket-based communication framework designed for real-time bidirectional messaging and state synchronization between multiple AI agents.
Mission Control
Mission Control is a CLI-first health aggregator that provides a unified view of agent processes, resource usage, and service health for autonomous AI infrastructure.
SkipUp Meeting Scheduler
An AI-powered scheduling assistant that automates multi-participant meeting coordination through intelligent, asynchronous email outreach.
Featured*
