OpenClaw

Offline Llama

offline llama

and-ray-m0 来源安装数☆ 0 仓库 Stars

查看来源 ↗安装与使用 SKILL.md ↗

安装与使用

请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求，再添加到工作流。

SkillHub CLI

skillhub install and-ray-m/offline-llama

打开来源仓库，查看说明和兼容环境后，按来源提供的命令或步骤安装。

将来源链接提供给你的 Agent，让它先读取说明、核对依赖，并向你说明将执行的安装步骤。

查看来源 ↗使用指南

技能说明

offline-llama

Autonomously manage and use local Ollama models for continuous operation without internet dependency. Includes model health monitoring, automatic fallback, and self-healing capabilities.

Overview

This skill enables autonomous operation with local Ollama models. It monitors model health, automatically switches between models when issues occur, and maintains functionality even without internet connectivity. The skill includes self-healing capabilities to restart services and clear resources when needed.

Core Features

Model Management

Health Monitoring: Continuously check model availability and performance
Automatic Fallback: Switch to alternative models when primary fails
Model Switching: Dynamically select best available model for task

Self-Healing

Service Restart: Automatically restart Ollama when models become unavailable
Resource Management: Clear cache and temporary files to free resources
Model Reinstallation: Reinstall problematic models automatically

Connectivity Awareness

Internet Detection: Monitor internet connectivity status
Smart Fallback: Switch to remote models when local models unavailable and internet is present
Offline Mode: Maintain full functionality without internet

Configuration

Models

Primary: llama-3.1-8b-instruct (general tasks)
Secondary: mistral-7b-instruct (faster responses)
Specialized: code-llama-7b (coding tasks)

Health Checks

Model Status: Monitor availability every 30 seconds
Latency Tracking: Monitor response times every minute
Resource Usage: Monitor GPU/CPU and memory every 5 minutes

Fallback Strategies

Model Switching: Automatically switch to alternative local models
Response Retry: Retry failed requests with exponential backoff
Degraded Mode: Continue with limited functionality if all models unavailable

Usage

When Internet is Available

Use local models...

来源信息

Source: github
Repository: openclaw/skills
Path: skills/and-ray-m/offline-llama
Version: —
Commit: 908bb3305b44a1217a69135a5a2c4c30974b7b35
分类来源: rules
状态: active
可见性: public

这些信息来自收录时的来源分析，可能滞后；使用前请核对最新来源文件。

来源内容

SKILL.md

# offline-llama

Autonomously manage and use local Ollama models for continuous operation without internet dependency. Includes model health monitoring, automatic fallback, and self-healing capabilities.

## Overview

This skill enables autonomous operation with local Ollama models. It monitors model health, automatically switches between models when issues occur, and maintains functionality even without internet connectivity. The skill includes self-healing capabilities to restart services and clear resources when needed.

## Core Features

### Model Management
- **Health Monitoring**: Continuously check model availability and performance
- **Automatic Fallback**: Switch to alternative models when primary fails
- **Model Switching**: Dynamically select best available model for task

### Self-Healing
- **Service Restart**: Automatically restart Ollama when models become unavailable
- **Resource Management**: Clear cache and temporary files to free resources
- **Model Reinstallation**: Reinstall problematic models automatically

### Connectivity Awareness
- **Internet Detection**: Monitor internet connectivity status
- **Smart Fallback**: Switch to remote models when local models unavailable and internet is present
- **Offline Mode**: Maintain full functionality without internet

## Configuration

### Models
- **Primary**: llama-3.1-8b-instruct (general tasks)
- **Secondary**: mistral-7b-instruct (faster responses)
- **Specialized**: code-llama-7b (coding tasks)

### Health Checks
- **Model Status**: Monitor availability every 30 seconds
- **Latency Tracking**: Monitor response times every minute
- **Resource Usage**: Monitor GPU/CPU and memory every 5 minutes

### Fallback Strategies
1. **Model Switching**: Automatically switch to alternative local models
2. **Response Retry**: Retry failed requests with exponential backoff
3. **Degraded Mode**: Continue with limited functionality if all models unavailable

## Usage

### When Internet is Available
- Use local models...

搜索