Menü
Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
npx skills add https://github.com/Aradotso/hermes-skills --skill awesome-openclaw-tutorialDer Befehl bleibt in einer Zeile. Scrollen Sie horizontal, um ihn vor dem Kopieren vollständig zu prüfen.
Sie bevorzugen eine lokale Kopie? Laden Sie die Dateien herunter, die SkillsMP derzeit vorliegen.
Basierend auf der SOC-Berufsklassifikation
SKILL.md wird angezeigt
| name | awesome-openclaw-tutorial |
| description | OpenClaw AI工作助手完整中文教程 - 安装配置、Skills扩展、自动化工作流和实战案例 |
| triggers | ["how do I install and configure OpenClaw","show me OpenClaw skills and automation examples","help me set up OpenClaw with custom models","how to use OpenClaw for file management and knowledge base","OpenClaw integration with Feishu Slack or WeChat","troubleshoot OpenClaw gateway or API connection issues","OpenClaw Task Flow and webhooks configuration","create OpenClaw workflows for productivity automation"] |
Skill by ara.so — Hermes Skills collection.
Complete Chinese tutorial for OpenClaw - an AI-powered work assistant. Covers installation, configuration, Skills ecosystem, automation workflows, API integration, and practical use cases for solo entrepreneurs and teams.
OpenClaw is an AI assistant framework that runs locally or in cloud, providing:
openclaw inferCurrent stable version: v2026.4.14
Recommended runtime: Node 24 (minimum Node 22.16+)
Feishu Miaoda (Free, 1M tokens/day):
Baota Panel (Free plugin, visual management):
# Install Baota Panel first
# Add "OpenClaw" plugin from marketplace
# One-click install and configure
# Install via npm (Node 22.16+ required)
npm install -g openclaw
# Or via yarn
yarn global add openclaw
# Initialize and onboard
openclaw onboard
# Start gateway
openclaw gateway start
# Pull latest image
docker pull openclaw/openclaw:latest
# Run with environment variables
docker run -d \
-e OPENCLAW_TZ=Asia/Shanghai \
-e OPENCLAW_API_KEY=$OPENCLAW_API_KEY \
-p 3000:3000 \
-v ~/.openclaw:/root/.openclaw \
openclaw/openclaw:latest
# Check status
docker ps
# Login to model providers (replaces old local-* skills)
openclaw models auth login --provider anthropic
openclaw models auth login --provider openai
openclaw models auth login --provider gemini
# List authenticated providers
openclaw models auth list
# Set default model
openclaw config set model.default "claude-3-5-sonnet-20241022"
# Enable fast mode (cheaper, faster for OpenAI/Anthropic)
# User says: /fast
# Set authentication mode (token or password)
openclaw config set gateway.auth.mode token
openclaw config set gateway.auth.token "${OPENCLAW_GATEWAY_TOKEN}"
# Restart gateway
openclaw gateway restart
# Set to 'full' for complete toolset including command execution
openclaw config set tools.profile full
# Profile options:
# - messaging: only chat and session management
# - default: standard tools (no command exec)
# - coding: programming tools
# - full: complete toolset (RECOMMENDED)
# - all: everything enabled
openclaw gateway restart
# Enable Active Memory (retrieves context before replies)
openclaw config set memory.active.enabled true
# Enable Dreaming (background memory consolidation)
openclaw config set memory.dreaming.enabled true
# Enable Memory Wiki (structured knowledge claims)
openclaw config set memory.wiki.enabled true
# Configure memory retention
openclaw config set memory.retention.days 90
# Check version
openclaw --version
# Show current config
openclaw config list
# Set config value
openclaw config set <key> <value>
# Gateway control
openclaw gateway start
openclaw gateway stop
openclaw gateway restart
openclaw gateway status
# Model inference (unified interface)
openclaw infer text "Explain quantum computing"
openclaw infer image "A red panda in space" --model dall-e-3
openclaw infer video "Ocean waves" --provider runway
openclaw infer audio "Calm piano music" --provider suno
openclaw infer tts "Hello world" --voice nova
openclaw infer web "https://example.com"
openclaw infer embedding "Document text to embed"
# Skills management (DEPRECATED in 2026.4+, use models auth instead)
# clawhub search <query>
# clawhub info <skill-name>
# clawhub install <skill-name>
# Create cron task (scheduled)
openclaw tasks create \
--name "daily-summary" \
--schedule "0 18 * * *" \
--command "Generate daily summary and send to Feishu"
# Create Task Flow (persistent workflow)
openclaw tasks flow create \
--name "content-pipeline" \
--trigger webhook \
--steps "research,write,review,publish"
# List tasks
openclaw tasks list
# Create webhook endpoint
openclaw webhooks create \
--name "github-pr-review" \
--url "/webhooks/github" \
--action "Review PR and comment"
# List webhooks
openclaw webhooks list
#!/bin/bash
# Auto-organize downloads folder
# User prompt to OpenClaw:
# "Organize my Downloads folder: move images to Images/,
# PDFs to Documents/, and delete files older than 30 days"
# OpenClaw executes internally via file management tools
# No manual scripting needed - AI handles the logic
# For custom automation, use Task Flow:
openclaw tasks flow create \
--name "organize-downloads" \
--schedule "0 2 * * *" \
--steps "scan_downloads,categorize_files,move_files,cleanup_old"
// Add document to knowledge base
const { OpenClawClient } = require('openclaw-sdk');
const client = new OpenClawClient({
apiKey: process.env.OPENCLAW_API_KEY,
baseUrl: 'http://localhost:3000'
});
async function addToKnowledgeBase(filePath, metadata) {
try {
const result = await client.knowledge.add({
file: filePath,
metadata: {
tags: metadata.tags,
source: metadata.source,
category: metadata.category
},
activeMemory: true, // Enable Active Memory indexing
memoryWiki: true // Add to Memory Wiki
});
console.log('Document added:', result.id);
return result;
} catch (error) {
console.error('Failed to add document:', error.message);
throw error;
}
}
// Query knowledge base
async function searchKnowledge(query) {
const results = await client.knowledge.search({
query: query,
limit: 10,
useActiveMemory: true, // Include Active Memory context
useMemoryWiki: true // Include Memory Wiki claims
});
return results;
}
// Example usage
addToKnowledgeBase('./report.pdf', {
tags: ['quarterly', 'finance'],
source: 'internal',
category: 'reports'
});
searchKnowledge('Q4 revenue analysis').then(results => {
console.log('Found:', results.length, 'documents');
});
// Feishu bot webhook handler
const express = require('express');
const app = express();
app.post('/feishu/webhook', async (req, res) => {
const event = req.body;
if (event.type === 'message') {
const userMessage = event.message.content;
// Send to OpenClaw for processing
const response = await fetch('http://localhost:3000/api/chat', {
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Authorization': `Bearer ${process.env.OPENCLAW_API_KEY}`
},
body: JSON.stringify({
message: userMessage,
sessionId: event.message.chat_id,
platform: 'feishu',
useActiveMemory: true // Retrieve relevant context
})
});
const result = await response.json();
// Send response back to Feishu
await sendFeishuMessage(event.message.chat_id, result.reply);
}
res.json({ success: true });
});
app.listen(3000);
#!/bin/bash
# Generate content pipeline: text -> image -> video
# Step 1: Generate article
ARTICLE=$(openclaw infer text "Write a 500-word article about sustainable energy" \
--model gpt-4-turbo)
# Step 2: Generate cover image
openclaw infer image "Sustainable energy concept art, vibrant, professional" \
--model dall-e-3 \
--size 1792x1024 \
--output cover.png
# Step 3: Generate promotional video
openclaw infer video "Solar panels in a green field, golden hour" \
--provider runway \
--duration 5 \
--output promo.mp4
# Step 4: Generate background music
openclaw infer audio "Uplifting corporate background music" \
--provider suno \
--duration 30 \
--output bgm.mp3
# Step 5: Combine and publish
echo "Content pipeline complete!"
ls -lh cover.png promo.mp4 bgm.mp3
# ~/.openclaw/flows/content-approval.yaml
name: content-approval-flow
trigger:
type: webhook
path: /webhooks/content-submit
steps:
- name: review-content
action: infer
input: "Review the submitted content for quality and compliance"
model: claude-3-5-sonnet-20241022
- name: check-approval
action: conditional
condition: "review.score > 8"
- name: publish
action: webhook
url: https://cms.example.com/publish
method: POST
headers:
Authorization: "Bearer ${CMS_API_KEY}"
body:
content: "{{ steps.review.content }}"
status: approved
- name: notify
action: send-message
platform: feishu
channel: "${FEISHU_CHANNEL_ID}"
message: "Content published: {{ steps.publish.url }}"
Load flow:
openclaw tasks flow load ~/.openclaw/flows/content-approval.yaml
openclaw tasks flow start content-approval-flow
# Create morning briefing task
openclaw tasks create \
--name "morning-briefing" \
--schedule "0 8 * * *" \
--command "Generate morning briefing: check emails, calendar, news, weather. Send to Feishu."
# Create evening summary task
openclaw tasks create \
--name "evening-summary" \
--schedule "0 18 * * *" \
--command "Summarize today's work: meetings attended, tasks completed, files created. Archive to knowledge base."
// Sync message across platforms
async function syncMessage(message, platforms) {
const results = await Promise.all(
platforms.map(platform =>
client.messaging.send({
platform: platform,
channelId: process.env[`${platform.toUpperCase()}_CHANNEL_ID`],
message: message,
useActiveMemory: true
})
)
);
return results;
}
// Usage
syncMessage('Deployment complete for v2.0', ['feishu', 'slack', 'wechat']);
# User uploads document
openclaw knowledge add ./research-paper.pdf \
--tags "AI,research" \
--auto-extract-claims
# OpenClaw automatically:
# 1. Extracts text content
# 2. Creates Memory Wiki claims with evidence
# 3. Indexes for Active Memory retrieval
# 4. Checks for contradictions with existing claims
# Query across knowledge base and memory
openclaw infer text "What are the latest findings on transformer models?" \
--use-active-memory \
--use-memory-wiki
Symptom: Error: gateway.auth.mode must be set
Fix:
openclaw config set gateway.auth.mode token
openclaw config set gateway.auth.token "$(openssl rand -hex 32)"
openclaw gateway restart
Symptom: OpenClaw responds to questions but won't manage files or run commands
Fix:
# Check current profile
openclaw config get tools.profile
# Set to 'full' for complete toolset
openclaw config set tools.profile full
openclaw gateway restart
Symptom: Model provider not authenticated
Fix:
# Re-authenticate (2026.4+ method)
openclaw models auth login --provider openai
openclaw models auth login --provider anthropic
# Verify
openclaw models auth list
# Set default
openclaw config set model.default "claude-3-5-sonnet-20241022"
Solutions:
# 1. Enable fast mode for OpenAI/Anthropic (cheaper)
openclaw config set model.fast_mode true
# 2. Use local models with Ollama
openclaw models auth login --provider ollama
openclaw config set model.default "llama3:70b"
# 3. Set token limits
openclaw config set model.max_tokens 4000
# 4. Use Chinese domestic models (95% cost savings)
openclaw models auth login --provider deepseek
openclaw config set model.default "deepseek-chat"
Symptom: AI doesn't recall previous context
Fix:
# Enable Active Memory
openclaw config set memory.active.enabled true
openclaw config set memory.dreaming.enabled true
# Force memory refresh
openclaw memory rebuild
# Check memory stats
openclaw memory stats
# Restart gateway
openclaw gateway restart
Symptom: Workflows stuck or fail silently
Debug:
# Check task logs
openclaw tasks logs <task-name>
# Validate flow definition
openclaw tasks flow validate ~/.openclaw/flows/my-flow.yaml
# Test individual steps
openclaw tasks flow test my-flow --step review-content
# Enable debug mode
openclaw config set debug.tasks true
openclaw gateway restart
Symptom: infer image fails with ComfyUI backend
Fix:
# Install ComfyUI skill (if using pre-2026.4 setup)
# Note: In 2026.4+, use built-in infer command instead
# Verify ComfyUI is running
curl http://localhost:8188/system_stats
# Set ComfyUI endpoint
openclaw config set comfyui.endpoint "http://localhost:8188"
# Test connection
openclaw infer image "test" --provider comfyui
Symptom: Black console window pops up on gateway restart
Fix: Upgrade to v2026.3.13+ which fixes this issue
# Update OpenClaw
npm update -g openclaw
# Verify version
openclaw --version # Should be >= v2026.3.13
Key config paths in ~/.openclaw/config.json:
{
"model": {
"default": "claude-3-5-sonnet-20241022",
"fast_mode": true,
"max_tokens": 4000
},
"gateway": {
"auth": {
"mode": "token",
"token": "${OPENCLAW_GATEWAY_TOKEN}"
},
"port": 3000
},
"tools": {
"profile": "full"
},
"memory": {
"active": {
"enabled": true
},
"dreaming": {
"enabled": true,
"schedule": "0 3 * * *"
},
"wiki": {
"enabled": true
},
"retention": {
"days": 90
}
},
"platforms": {
"feishu": {
"app_id": "${FEISHU_APP_ID}",
"app_secret": "${FEISHU_APP_SECRET}"
},
"slack": {
"bot_token": "${SLACK_BOT_TOKEN}"
}
}
}
openclaw models auth (2026.4+) or legacy clawhub searchImportant: This tutorial tracks v2026.4.14 stable. Some legacy content (old skill names, deprecated commands) remains for reference. Always verify against current openclaw --version and official docs.