Coala Client for Openclaw

A versatile command-line interface for interacting with LLMs and managing Model Context Protocol (MCP) servers and AI skills.

hubentu
v0.1.2
Feb 20, 2026
0
1.6k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install coala

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install coala 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 Coala Client?

Coala Client is a foundational component of the coala ecosystem, providing a streamlined CLI for developers to interact with OpenAI-compatible LLMs including OpenAI, Gemini, and Ollama. Its primary value lies in its deep integration with the Model Context Protocol (MCP), allowing users to bridge the gap between conversational AI and local or remote tools. By leveraging Openclaw Skills, Coala Client enables the seamless import of CWL toolsets and modular logic, transforming a standard chat interface into a powerful, extensible AI agent environment.

Coala Client Use Cases

  • Running interactive chat sessions that utilize local MCP tools for real-world tasks.
  • Executing single-prompt queries via the CLI with full access to registered toolsets.
  • Importing and managing CWL (Common Workflow Language) toolsets as functional MCP servers.
  • Loading modular Openclaw Skills from remote GitHub repositories or local directories to extend agent capabilities.

How Coala Client Works

  1. The user initializes the local environment which creates necessary configuration directories for MCP servers and skills.
  2. API keys for providers like OpenAI or Gemini are configured via environment variables or a local env file.
  3. The client imports MCP servers or CWL toolsets, registering them in a central JSON configuration for tool discovery.
  4. During a chat session, the LLM analyzes user requests and identifies when to call registered MCP tools to fetch data or perform actions.
  5. Users can dynamically inject context-specific logic by loading installed Openclaw Skills directly into the active chat session.

Coala Client Setup

Install the client and initialize the configuration directory:

uv tool install coala-client
coala init

Set your desired provider API key and start a session:

export OPENAI_API_KEY="your_api_key"
coala chat

To import new Openclaw Skills from a source:

coala skill https://github.com/owner/repo/tree/main/skills

Coala Client Data Schema & Taxonomy

Coala Client organizes its operations within the ~/.config/coala/ directory using the following structure:

Path Purpose Content
mcps/mcp_servers.json MCP Registry Definitions and connection strings for all MCP servers.
mcps/<toolset>/ Toolset Data Specific directories containing run_mcp.py and CWL definitions.
skills/ Skills Library Subfolders containing imported Openclaw Skills such as SKILL.md files.
env Environment Local storage for provider URLs and sensitive API keys.

Coala Client Advanced Features

  • Direct MCP tool execution via mcp-call for testing specific functions without an LLM loop.
  • Dynamic provider switching during active sessions using the /switch command.
  • Multi-source skill importing supporting local directories, ZIP files, and GitHub tree URLs.
  • Granular MCP control allowing users to disable all tools or specific servers to manage context window usage.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*