AgentClaw

https://github.com/wuding129/agentclaw
Admin
AgentClaw is a multi-tenant AI agent platform that upgrades OpenClaw single-user agents into a...
Tokens:462,969
Snippets:5,432
Trust Score:6.9
License:MIT
Update:5 months ago
Show doc for...
Tokens:
Raw
### Summarize Content Examples

Source: https://github.com/wuding129/agentclaw/blob/main/skills/summarize/SKILL.md

Quick start examples for summarizing different types of content. Ensure the 'summarize' binary is installed.

```bash
summarize "https://example.com" --model google/gemini-3-flash-preview
```

```bash
summarize "/path/to/file.pdf" --model google/gemini-3-flash-preview
```

```bash
summarize "https://youtu.be/dQw4w9WgXcQ" --youtube auto
```

--------------------------------

### Install and Start Frontend Development Server

Source: https://context7.com/wuding129/agentclaw/llms.txt

Set up the frontend by navigating to the directory, installing Node.js dependencies, and starting the Vite development server.

```bash
cd frontend
npm install
npm run dev
```

--------------------------------

### Install the Go SDK

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/go/claude-api.md

Use 'go get' to install the anthropic-sdk-go package.

```bash
go get github.com/anthropics/anthropic-sdk-go
```

--------------------------------

### Multi-platform Setup Example

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/configuration-examples.md

Configures agents with a workspace and enables multiple communication channels like WhatsApp, Telegram, and Discord.

```json5
{
  agent: { workspace: "~/.openclaw/workspace" },
  channels: {
    whatsapp: { allowFrom: ["+15555550123"] },
    telegram: {
      enabled: true,
      botToken: "YOUR_TOKEN",
      allowFrom: ["123456789"],
    },
    discord: {
      enabled: true,
      token: "YOUR_TOKEN",
      dm: { allowFrom: ["123456789012345678"] },
    },
  },
}
```

--------------------------------

### Setup Environment Variables

Source: https://github.com/wuding129/agentclaw/blob/main/README.md

Copy the example environment file and edit it to add your LLM API keys. This is a required step before running the application.

```bash
cp .env.example .env
# edit .env and add LLM API keys
```

--------------------------------

### Quick Start with Agent SDK

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/typescript/agent-sdk/README.md

Begin using the Agent SDK by importing the `query` function. This example demonstrates a basic query to explain a codebase, allowing specific tools.

```typescript
import { query } from "@anthropic-ai/claude-agent-sdk";

for await (const message of query({
  prompt: "Explain this codebase",
  options: { allowedTools: ["Read", "Glob", "Grep"] },
})) {
  if ("result" in message) {
    console.log(message.result);
  }
}
```

--------------------------------

### Install and Start Bridge Service

Source: https://context7.com/wuding129/agentclaw/llms.txt

Navigate to the bridge directory, install dependencies, and start the bridge service using TypeScript/Node.js. This compiles and starts the bridge, spawning the OpenClaw Gateway.

```bash
cd bridge
npm install
tsx start.ts
```

--------------------------------

### Complete Python MCP Server Example

Source: https://github.com/wuding129/agentclaw/blob/main/skills/mcp-builder/reference/python_mcp_server.md

This example demonstrates the setup of a complete Python MCP server, including necessary imports, initialization, constants, enums, and Pydantic models for input validation.

```python
#!/usr/bin/env python3
'''
MCP Server for Example Service.

This server provides tools to interact with Example API, including user search,
project management, and data export capabilities.
'''

from typing import Optional, List, Dict, Any
from enum import Enum
import httpx
from pydantic import BaseModel, Field, field_validator, ConfigDict
from mcp.server.fastmcp import FastMCP

# Initialize the MCP server
mcp = FastMCP("example_mcp")

# Constants
API_BASE_URL = "https://api.example.com/v1"

# Enums
class ResponseFormat(str, Enum):
    '''Output format for tool responses.'''
    MARKDOWN = "markdown"
    JSON = "json"

# Pydantic Models for Input Validation
class UserSearchInput(BaseModel):
    '''Input model for user search operations.'''
    model_config = ConfigDict(
        str_strip_whitespace=True,
        validate_assignment=True
    )

    query: str = Field(..., description="Search string to match against names/emails", min_length=2, max_length=200)
    limit: Optional[int] = Field(default=20, description="Maximum results to return", ge=1, le=100)
    offset: Optional[int] = Field(default=0, description="Number of results to skip for pagination", ge=0)
    response_format: ResponseFormat = Field(default=ResponseFormat.MARKDOWN, description="Output format")

    @field_validator('query')
    @classmethod
    def validate_query(cls, v: str) -> str:
        if not v.strip():
            raise ValueError("Query cannot be empty or whitespace only")
        return v.strip()
```

--------------------------------

### Quick Start Streaming

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/typescript/claude-api/streaming.md

A basic example demonstrating how to initiate a stream and process text deltas.

```APIDOC
## Quick Start Streaming

### Description
This snippet shows the fundamental way to start a streaming request and iterate over the incoming events to process text content.

### Method
`client.messages.stream()`

### Parameters
- `model` (string) - Required - The model to use for the request.
- `max_tokens` (number) - Required - The maximum number of tokens to generate.
- `messages` (Array<object>) - Required - An array of message objects representing the conversation history.

### Request Example
```typescript
const stream = client.messages.stream({
  model: "claude-opus-4-6",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Write a story" }],
});

for await (const event of stream) {
  if (
    event.type === "content_block_delta" &&
    event.delta.type === "text_delta"
  ) {
    process.stdout.write(event.delta.text);
  }
}
```

### Response
Events are streamed incrementally. The `content_block_delta` with `text_delta` type provides the text chunks.
```

--------------------------------

### Sandbox Script Execution Example

Source: https://github.com/wuding129/agentclaw/blob/main/README.md

Demonstrates commands for installing dependencies and running a Python script within the sandbox environment. Includes setting a timeout for execution.

```bash
# install deps
apt-get update && apt-get install -y some-package
pip install requests

# run test
timeout 30 python3 scripts/main.py
```

--------------------------------

### Recommended OpenClaw Installation

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Installs OpenClaw using the recommended method and onboards the system, setting up the gateway.

```bash
curl -fsSL https://openclaw.ai/install.sh | bash
```

```bash
openclaw onboard --install-daemon
```

--------------------------------

### Setup Anthropic Token and Check Model Status

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/concepts/models.md

This example demonstrates setting up an Anthropic token and then checking the model status.

```bash
claude setup-token
openclaw models status
```

--------------------------------

### Start Local Development Services

Source: https://github.com/wuding129/agentclaw/blob/main/CLAUDE.md

Use this command to start all necessary services for local development, including PostgreSQL, Bridge, Gateway, and Frontend. You can also specify services to start or skip.

```bash
python start_local.py

# Start specific services only
python start_local.py --only db,gateway,frontend

# Skip specific services
python start_local.py --skip bridge

# Stop all services
python start_local.py --stop
```

--------------------------------

### Install and Connect Tailscale on VPS

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Steps to install Tailscale on a Virtual Private Server (VPS) and connect from a Mac. This involves installing the Tailscale client on both machines and logging into the same tailnet.

```bash
curl -fsSL https://tailscale.com/install.sh | sh
sudo tailscale up
```

--------------------------------

### Start Gateway Locally

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/index.md

Use these commands to start the Gateway service on your local machine. The `--verbose` flag mirrors debug/trace output to stdio. `--force` can be used to kill an existing listener on the selected port before starting.

```bash
openclaw gateway --port 18789
# debug/trace mirrored to stdio
openclaw gateway --port 18789 --verbose
# force-kill listener on selected port, then start
openclaw gateway --force
```

--------------------------------

### Onboard and Install OpenClaw Service

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/index.md

Initializes OpenClaw and installs it as a background service on your system.

```bash
openclaw onboard --install-daemon
```

--------------------------------

### Install Claude Agent SDK

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/python/agent-sdk/README.md

Install the SDK using pip. This is the first step to using the agent capabilities.

```bash
pip install claude-agent-sdk
```

--------------------------------

### Re-run Installer with Verbose Output (Linux/macOS)

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Use this command to get detailed output from the OpenClaw installer, which is helpful for diagnosing installation issues. This applies to the standard installer.

```bash
curl -fsSL https://openclaw.ai/install.sh | bash -s -- --verbose
```

--------------------------------

### Environment Variable Setup

Source: https://github.com/wuding129/agentclaw/blob/main/CLAUDE.md

Create a `.env` file in the project root to configure environment variables. Refer to `.env.example` for available options.

```bash
# Example .env content
# OPENAI_API_KEY=your_openai_api_key
# DATABASE_URL=your_database_url
```

--------------------------------

### Install OpenClaw from Git Checkout

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Installs OpenClaw from a git checkout, allowing AI agents to read code and docs. Use this for local setup and debugging.

```bash
curl -fsSL https://openclaw.ai/install.sh | bash -s -- --install-method git
```

--------------------------------

### One-time DNS Server Setup for Gateway

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/bonjour.md

Installs and configures CoreDNS on the gateway host to serve DNS-SD records for a chosen domain over Tailscale. This is a one-time setup command.

```bash
openclaw dns setup --apply
```

--------------------------------

### Full CLI backend configuration example

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/cli-backends.md

This example shows comprehensive configuration options for a custom CLI backend, including arguments, output parsing, session handling, and image support.

```json5
{
  agents: {
    defaults: {
      cliBackends: {
        "claude-cli": {
          command: "/opt/homebrew/bin/claude",
        },
        "my-cli": {
          command: "my-cli",
          args: ["--json"],
          output: "json",
          input: "arg",
          modelArg: "--model",
          modelAliases: {
            "claude-opus-4-6": "opus",
            "claude-opus-4-5": "opus",
            "claude-sonnet-4-5": "sonnet",
          },
          sessionArg: "--session",
          sessionMode: "existing",
          sessionIdFields: ["session_id", "conversation_id"],
          systemPromptArg: "--system",
          systemPromptWhen: "first",
          imageArg: "--image",
          imageMode: "repeat",
          serialize: true,
        },
      },
    },
  },
}
```

--------------------------------

### Quick Start with `query()`

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/python/agent-sdk/README.md

Demonstrates the simplest way to use the agent SDK with the `query` function for one-shot interactions. It shows how to specify allowed tools and process the results.

```APIDOC
## `query()` — Simple One-Shot Usage

The `query()` function is the simplest way to run an agent. It returns an async iterator of messages.

### Usage Example
```python
from claude_agent_sdk import query, ClaudeAgentOptions, ResultMessage

async for message in query(
    prompt="Explain this codebase",
    options=ClaudeAgentOptions(allowed_tools=["Read", "Glob", "Grep"])
):
    if isinstance(message, ResultMessage):
        print(message.result)
```
```

--------------------------------

### Message Request with System Prompt

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/python/claude-api/README.md

Send a message request with a system prompt to guide the AI's behavior. This example specifies that the assistant should provide Python examples.

```python
response = client.messages.create(
    model="claude-opus-4-6",
    max_tokens=1024,
    system="You are a helpful coding assistant. Always provide examples in Python.",
    messages=[{"role": "user", "content": "How do I read a JSON file?"}]
)
```

--------------------------------

### Start Local AgentClaw Services

Source: https://context7.com/wuding129/agentclaw/llms.txt

Use these commands to start or stop the local AgentClaw environment. The `--skip bridge` option allows using an existing OpenClaw installation.

```bash
python start_local.py --skip bridge
```

```bash
python start_local.py --stop
```

--------------------------------

### Onboard Anthropic with Setup Token

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/concepts/model-providers.md

CLI command to onboard the Anthropic provider using a setup-token. Paste the token when prompted.

```bash
openclaw onboard --auth-choice token
```

--------------------------------

### Install OpenClaw Chrome Extension

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Commands to install the OpenClaw browser extension and get its path for manual loading in Chrome. Ensure Developer mode is enabled in Chrome extensions.

```bash
openclaw browser extension install
```

```bash
openclaw browser extension path
```

--------------------------------

### Run OpenClaw Onboarding Wizard

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/concepts/models.md

Use the onboard command to set up model and authentication for common providers.

```bash
openclaw onboard
```

--------------------------------

### Full End-to-End Example

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/python/claude-api/batches.md

A comprehensive example demonstrating the entire workflow of creating a batch, waiting for its completion, and collecting the results.

```APIDOC
## Full End-to-End Example

A comprehensive example demonstrating the entire workflow of creating a batch, waiting for its completion, and collecting the results.

### Workflow Steps:
1. **Prepare requests**: Define the individual message requests, each with a unique `custom_id` and `params`.
2. **Create batch**: Submit the list of requests to the `/v1/messages/batches` endpoint.
3. **Wait for completion**: Poll the batch status using the retrieve endpoint until `processing_status` is 'ended'.
4. **Collect results**: Iterate through the batch results using the results endpoint, processing succeeded or errored requests.
```

--------------------------------

### Client Initialization

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/python/claude-api/README.md

Demonstrates how to initialize the Anthropic client, both with and without an explicit API key, and how to set up an asynchronous client.

```APIDOC
## Client Initialization

```python
import anthropic

# Default (uses ANTHROPIC_API_KEY env var)
client = anthropic.Anthropic()

# Explicit API key
client = anthropic.Anthropic(api_key="your-api-key")

# Async client
async_client = anthropic.AsyncAnthropic()
```
```

--------------------------------

### OpenClaw Multi-Instance Setup with Profiles

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Demonstrates how to set up and run multiple OpenClaw Gateway instances on the same host using profiles. This involves isolating configuration, state, and ports for each instance.

```bash
openclaw --profile <name> gateway install
```

--------------------------------

### Message Request with System Prompt

Source: https://github.com/wuding129/agentclaw/blob/main/skills/claude-api/typescript/claude-api/README.md

Include a system prompt to guide the AI's behavior, such as instructing it to provide Python examples.

```typescript
const response = await client.messages.create({
  model: "claude-opus-4-6",
  max_tokens: 1024,
  system:
    "You are a helpful coding assistant. Always provide examples in Python.",
  messages: [{ role: "user", content: "How do I read a JSON file?" }],
});
```

--------------------------------

### Example Crop for Name Field with ImageMagick

Source: https://github.com/wuding129/agentclaw/blob/main/skills/pdf/forms.md

This is a specific example of using ImageMagick to crop an image for refining the coordinates of a 'Name' field. The crop region is set to 300x80 pixels, starting at offset (50, 120) from the top-left corner of the page image.

```bash
magick images_dir/page_1.png -crop 300x80+50+120 +repage crops/name_field.png
```

--------------------------------

### Install Gateway Services with Profiles

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/multiple-gateways.md

Install the Gateway as a service for each profile. This ensures that each service instance runs with its designated configuration and state.

```bash
openclaw --profile main gateway install
openclaw --profile rescue gateway install
```

--------------------------------

### Configure and Run Platform Gateway

Source: https://context7.com/wuding129/agentclaw/llms.txt

Set up the Platform Gateway by installing dependencies, configuring environment variables for database and gateway URLs, and running the FastAPI application with Uvicorn for development.

```bash
cd platform
pip install -e .[dev]
export PLATFORM_DATABASE_URL="postgresql+asyncpg://frameclaw:frameclaw@localhost:5432/frameclaw_platform"
export PLATFORM_DEV_OPENCLAW_URL="http://127.0.0.1:18080"
export PLATFORM_DEV_GATEWAY_URL="ws://127.0.0.1:18789"
python -m uvicorn app.main:app --host 0.0.0.0 --port 8080 --reload
```

--------------------------------

### Caddy Reverse Proxy Configuration

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/channels/googlechat.md

Example Caddyfile configuration to proxy only the /googlechat path to your OpenClaw gateway. This setup ignores other paths, enhancing security.

```caddy
your-domain.com {
    reverse_proxy /googlechat* localhost:18789
}
```

--------------------------------

### Example TOOLS.md Configuration

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/reference/templates/TOOLS.md

This markdown snippet demonstrates how to structure environment-specific notes for tools, including camera names and locations, SSH host aliases, and preferred TTS settings. This file is intended for personal configuration details that should not be shared as part of general skills.

```markdown
### Cameras

- living-room → Main area, 180° wide angle
- front-door → Entrance, motion-triggered

### SSH

- home-server → 192.168.1.100, user: admin

### TTS

- Preferred voice: "Nova" (warm, slightly British)
- Default speaker: Kitchen HomePod
```

--------------------------------

### Setup and Run Gateways with Profiles

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/multiple-gateways.md

Use profiles to manage separate configurations and state directories for each Gateway instance. Ensure unique ports are assigned to avoid conflicts.

```bash
# main
openclaw --profile main setup
openclaw --profile main gateway --port 18789

# rescue
openclaw --profile rescue setup
openclaw --profile rescue gateway --port 19001
```

--------------------------------

### Run OpenClaw Bridge (TypeScript/Node.js)

Source: https://github.com/wuding129/agentclaw/blob/main/CLAUDE.md

Commands for installing dependencies and starting the OpenClaw Bridge in development mode. Also includes instructions for building the TypeScript code.

```bash
cd bridge
npm install
tsx start.ts     # Start bridge + OpenClaw Gateway (dev mode)
npx tsc          # Build bridge TypeScript
```

--------------------------------

### Install from marketplace

Source: https://github.com/wuding129/agentclaw/blob/main/bridge/skills/platform/SKILL.md

Installs a skill from the public marketplace using its slug. The slug typically follows the format 'owner/repo@skill-name'.

```APIDOC
## Install from marketplace

### Description
Installs a skill from the public marketplace using its slug. The slug typically follows the format 'owner/repo@skill-name'.

### Method
POST

### Endpoint
/api/marketplaces/skills/install

### Parameters
#### Request Body
- **slug** (string) - Required - The unique slug of the skill to install (e.g., "owner/repo@skill-name").
```

--------------------------------

### OpenClaw Absolute Minimum Configuration

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/configuration-examples.md

This is the most basic configuration for OpenClaw, essential for getting started. Save this to `~/.openclaw/openclaw.json` to enable direct messaging with the bot from a specified WhatsApp number.

```json5
{
  agent: { workspace: "~/.openclaw/workspace" },
  channels: { whatsapp: { allowFrom: ["+15555550123"] } },
}
```

--------------------------------

### OpenClaw Gateway CLI with Tailscale

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/tailscale.md

Command-line interface examples for starting the OpenClaw Gateway with Tailscale Serve or Funnel modes. The `--auth password` flag is used for Funnel mode.

```bash
openclaw gateway --tailscale serve
```

```bash
openclaw gateway --tailscale funnel --auth password
```

--------------------------------

### Run Platform Gateway (FastAPI)

Source: https://github.com/wuding129/agentclaw/blob/main/CLAUDE.md

Instructions for setting up and running the Platform Gateway locally. Ensure the database URL is set in your environment variables before running.

```bash
cd platform
# Install dependencies
pip install -e .[dev]

# Run with auto-reload
export PLATFORM_DATABASE_URL="postgresql+asyncpg://frameclaw:frameclaw@localhost:5432/frameclaw_platform"
python -m uvicorn app.main:app --host 0.0.0.0 --port 8080 --reload

# Run tests
pytest
```

--------------------------------

### Onboard Google Gemini with API Key

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/concepts/model-providers.md

CLI command to onboard the Google Gemini provider using a GEMINI_API_KEY.

```bash
openclaw onboard --auth-choice gemini-api-key
```

--------------------------------

### Example Error Handling in TypeScript

Source: https://github.com/wuding129/agentclaw/blob/main/skills/mcp-builder/reference/mcp_best_practices.md

Demonstrates how to catch errors during an operation and return a user-friendly error message, suggesting a potential solution. This pattern helps prevent exposing internal error details while guiding the user.

```typescript
try {
  const result = performOperation();
  return { content: [{ type: "text", text: result }] };
} catch (error) {
  return {
    isError: true,
    content: [{
      type: "text",
      text: `Error: ${error.message}. Try using filter='active_only' to reduce results.`
    }]
  };
}
```

--------------------------------

### Install and verify imsg CLI

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/channels/imessage.md

Installs the imsg CLI using Homebrew and verifies its installation by running the help command. Ensure imsg is installed before configuring OpenClaw.

```bash
brew install steipete/tap/imsg
imsg rpc --help
```

--------------------------------

### Initialize and Render

Source: https://github.com/wuding129/agentclaw/blob/main/skills/skill-creator/eval-viewer/viewer.html

Initializes the application and renders the benchmark view. This is the main entry point for the viewer functionality.

```javascript
container.innerHTML = html; } // ---- Start ---- init(); renderBenchmark();
```

--------------------------------

### Install Nostr Plugin

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/channels/nostr.md

Install the Nostr plugin for OpenClaw using npm. Restart the Gateway after installation.

```bash
openclaw plugins install @openclaw/nostr
```

--------------------------------

### Install OpenClaw Beta Version (macOS/Linux)

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Use this command to install the beta version of OpenClaw. Ensure you have curl installed.

```bash
curl -fsSL --proto '=https' --tlsv1.2 https://openclaw.ai/install.sh | bash -s -- --beta
```

--------------------------------

### Initialize Web Artifact Project

Source: https://github.com/wuding129/agentclaw/blob/main/skills/web-artifacts-builder/SKILL.md

Run this script to set up a new React project for building web artifacts. It configures Vite, Tailwind CSS, shadcn/ui, and Parcel for bundling.

```bash
bash scripts/init-artifact.sh <project-name>
cd <project-name>
```

--------------------------------

### Install ClawHub CLI with pnpm

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Install the ClawHub command-line interface using pnpm. This is an alternative method for installing skills on Linux.

```bash
pnpm add -g clawhub
```

--------------------------------

### Install Nostr Plugin with Local Checkout

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/channels/nostr.md

Install the Nostr plugin using a local checkout for development workflows. Restart the Gateway after installation.

```bash
openclaw plugins install --link <path-to-openclaw>/extensions/nostr
```

--------------------------------

### Configure GCP Project and Enable APIs

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/automation/gmail-pubsub.md

Log in to GCP, set the project, and enable the Gmail and Pub/Sub APIs.

```bash
gcloud auth login
gcloud config set project <project-id>
```

```bash
gcloud services enable gmail.googleapis.com pubsub.googleapis.com
```

--------------------------------

### Re-run Beta Installer with Verbose Output (Linux/macOS)

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Installs the beta version of OpenClaw with verbose logging enabled to help troubleshoot installation problems.

```bash
curl -fsSL https://openclaw.ai/install.sh | bash -s -- --beta --verbose
```

--------------------------------

### List Curated Skills with Install Status

Source: https://context7.com/wuding129/agentclaw/llms.txt

Returns all admin-curated skills with an `installed` flag. Use this to display available skills and their installation status to the user.

```bash
curl -s http://localhost:8084/api/skills/curated \
  -H "Authorization: Bearer $TOKEN" | jq
```

--------------------------------

### Install and Update Skills with ClawHub

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Use ClawHub to install new skills or update all existing ones. Skills are installed into the local `./skills` directory by default.

```bash
clawhub install <skill-slug>
```

```bash
clawhub update --all
```

--------------------------------

### Install OpenClaw from Source for Development

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/help/faq.md

Clones the OpenClaw repository, installs dependencies, builds the project and UI, and then onboards the system. Suitable for contributors and developers.

```bash
git clone https://github.com/openclaw/openclaw.git
```

```bash
cd openclaw
```

```bash
pnpm install
```

```bash
pnpm build
```

```bash
pnpm ui:build # auto-installs UI deps on first run
```

```bash
openclaw onboard
```

--------------------------------

### Install a marketplace skill

Source: https://github.com/wuding129/agentclaw/blob/main/bridge/skills/platform/SKILL.md

Installs a skill from the marketplace using its slug. The slug format is typically 'owner/repo@skill-name'.

```bash
curl -s -X POST http://localhost:18080/api/marketplaces/skills/install \
  -H "Content-Type: application/json" \
  -d '{"slug": "owner/repo@skill-name"}' \
  | python3 -m json.tool
```

```bash
npx --yes skills add owner/repo@skill-name -g -y --copy
```

--------------------------------

### Non-streaming Request Example

Source: https://github.com/wuding129/agentclaw/blob/main/docs/openclaw/docs/gateway/openresponses-http-api.md

Example of a non-streaming request to the /v1/responses endpoint.

```APIDOC
## POST /v1/responses

### Description
Sends a request to the OpenResponses API for a non-streaming response.

### Method
POST

### Endpoint
/v1/responses

### Request Example
```bash
curl -sS http://127.0.0.1:18789/v1/responses \
  -H 'Authorization: Bearer YOUR_TOKEN' \
  -H 'Content-Type: application/json' \
  -H 'x-openclaw-agent-id: main' \
  -d '{
    "model": "openclaw",
    "input": "hi"
  }'
```
```

--------------------------------

### Skill File Structure Example

Source: https://github.com/wuding129/agentclaw/blob/main/skills/skill-creator/SKILL.md

Illustrates the directory structure for a skill, including the SKILL.md file and optional bundled resources like scripts, references, and assets.

```tree
skill-name/
├── SKILL.md (required)
│   ├── YAML frontmatter (name, description required)
│   └── Markdown instructions
└── Bundled Resources (optional)
    ├── scripts/    - Executable code for deterministic/repetitive tasks
    ├── references/ - Docs loaded into context as needed
    └── assets/     - Files used in output (templates, icons, fonts)
```

--------------------------------

### List installed skills

Source: https://github.com/wuding129/agentclaw/blob/main/bridge/skills/platform/SKILL.md

Retrieves a list of all skills currently installed in the container.

```APIDOC
## List installed skills

### Description
Retrieves a list of all skills currently installed in the container.

### Method
GET

### Endpoint
/api/skills
```

--------------------------------

### Initialize Presentation and Add Basic Text

Source: https://github.com/wuding129/agentclaw/blob/main/skills/pptx/pptxgenjs.md

Sets up a new presentation with a specified layout and author, then adds a simple text slide. Ensure 'pptxgenjs' is installed and required.

```javascript
const pptxgen = require("pptxgenjs");

let pres = new pptxgen();
pres.layout = 'LAYOUT_16x9';  // or 'LAYOUT_16x10', 'LAYOUT_4x3', 'LAYOUT_WIDE'
pres.author = 'Your Name';
pres.title = 'Presentation Title';

let slide = pres.addSlide();
slide.addText("Hello World!", { x: 0.5, y: 0.5, fontSize: 36, color: "363636" });

pres.writeFile({ fileName: "Presentation.pptx" });
```

--------------------------------

### List installed skills

Source: https://github.com/wuding129/agentclaw/blob/main/bridge/skills/platform/SKILL.md

Retrieves a list of all skills currently installed in the environment.

```bash
curl -s http://localhost:18080/api/skills | python3 -m json.tool
```
Tokens:
Raw