1. Conduid
  2. AI
  3. Dojo
MCP server · AI

Dojo

MCP Dojo — a practice space for the Model Context Protocol: explore MCP servers hands-on, watch every exchange, replay it anywhere

Unclaimed ai
37Low

Scored 13 days ago · breakdown

About Dojo

Dojo is an MCP server in the AI category: mCP Dojo — a practice space for the Model Context Protocol: explore MCP servers hands-on, watch every exchange, replay it anywhere.

The repository has 5 stars and 0 forks.

Install

Claude Code
claude mcp add dojo -- npx -y @opensourcewtf/dojo-mcp
npx
npx -y @opensourcewtf/dojo-mcp

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Ask AI

✦

Ask AI about Dojo

Powered by Claude · Grounded in docs

I know everything about Dojo. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

  • ✓README presentThe repository ships documentation.

5 checks have not been run yet.

README


⚡ Quick Start

# Install globally
npm install -g @opensourcewtf/dojo

# Search for skills
dojo search testing

# Install a skill
dojo learn test-generation

# Install an MCP server
dojo learn --mcp playwright

✨ Features

Feature Description
🔍 Search Fuzzy search across 80+ skills from the curated registry
📦 Multi-Agent Install One command installs to Claude, Gemini, Cursor, and Codex
🔌 MCP Server Setup Auto-configures MCP servers with env var prompting
🔗 Skill Syncing Keep skills consistent across all your AI agents
⚡ CLI Detection Automatically detects which agents you have installed
🛡️ Registry Isolation Caches are namespaced by registry URL to prevent conflicts

🤖 Supported Agents (21)

Dojo supports 21 AI coding assistants across three categories:

Core Agents

Agent Skill Directory Format MCP Config
Claude .claude/skills/{skill}/SKILL.md folder-skill ~/.claude.json
Gemini .gemini/skills/{skill}/SKILL.md folder-skill ~/.gemini/settings.json
Antigravity .agent/workflows/{skill}.md flat-md ~/.gemini/antigravity/mcp_config.json
Cursor .cursor/rules/{skill}/RULE.md folder-rule ~/.cursor/mcp.json
Codex .codex/skills/{skill}/SKILL.md folder-skill —

CAM-Compatible Agents

These agents are compatible with block/ai-rules:

Agent Skill Directory Format MCP Config Docs
Amp .agents/skills/{skill}/SKILL.md folder-skill — Sourcegraph Amp
Cline .clinerules/{skill}.md flat-md (UI Settings) docs.cline.bot
Copilot .github/copilot-instructions/{skill}.md flat-md — GitHub Copilot
Firebender .firebender/skills/{skill}/SKILL.md folder-skill — Firebender
Goose .goose/skills/{skill}.md flat-md — Block/Square Goose
Kilocode .kilocode/rules/{skill}.md flat-md — Kilocode
Roo .roo/rules/{skill}.md flat-md .roo/mcp.json Roo

Dojo-Exclusive Agents

Not supported by CAM—unique to Dojo:

Agent Skill Directory Format MCP Config Docs
Windsurf .windsurf/workflows/{skill}.md flat-md ~/.codeium/windsurf/mcp_config.json docs.windsurf.com
Aider .aider/skills/{skill}.md flat-md — aider.chat
Zed AI .zed/skills/{skill}.md flat-md ~/.config/zed/settings.json zed.dev/docs/ai/rules
Cody .cody/skills/{skill}.md flat-md — Sourcegraph Cody
Void .void/skills/{skill}.md flat-md mcp_config.json Void Editor
Junie .junie/skills/{skill}.md flat-md (UI Settings) jetbrains.com/help/junie
Trae .trae/skills/{skill}.md flat-md — ByteDance Trae
Bolt .bolt/skills/{skill}.md flat-md — StackBlitz Bolt
Lovable .lovable/skills/{skill}.md flat-md — Lovable AI

Skill Formats

Format Structure Used By
folder-skill {dir}/{skill}/SKILL.md Claude, Gemini, Codex, Amp, Firebender
folder-rule {dir}/{skill}/RULE.md + frontmatter Cursor
flat-md {dir}/{skill}.md All others

Detection: Dojo automatically detects installed agents by checking if their CLI is in your PATH.


📖 CLI Reference

dojo learn <skill> (alias: add)

Install a skill or MCP server to all detected agents.

dojo learn test-generation              # Install a skill
dojo learn --mcp playwright             # Install MCP server (auto-finds mcp-playwright)
dojo learn mcp-github --for claude      # Install only to Claude
dojo learn skill-name -g                # Install globally to ~/.dojo/skills

Options:

Option Description
-g, --global Install to global ~/.dojo/skills (shared across projects)
--mcp Install MCP servers only (skip skill files)
--for <agents> Target specific agents (comma-separated: claude,gemini,cursor,codex)
--registry <url> Custom registry (local path, github:owner/repo, or full URL)

Modal Behavior: With --mcp, Dojo filters search results to only show skills with MCP servers and automatically finds mcp-<name> variants.


dojo search <query>

Search the skill registry. By default shows only skill files (excludes MCP-only entries).

dojo search testing             # Find testing skills
dojo search playwright --mcp    # Find MCP servers only
dojo search @anthropics         # Find skills by organization

Options:

Option Description
--mcp Show only MCP servers (modal filter)
--registry <url> Custom registry URL

dojo list (alias: ls)

Show installed skills or configured MCP servers.

dojo list                       # List installed skill files
dojo list --mcp                 # List configured MCP servers

Options:

Option Description
--mcp Show configured MCP servers instead of skills

dojo unlearn <skill> (alias: rm)

Remove a skill from agent directories.

dojo unlearn test-generation           # Remove skill from project
dojo unlearn skill-name -g             # Remove globally
dojo unlearn playwright --mcp          # Remove MCP config only (keep skill files)
dojo unlearn debugging -y              # Skip confirmation

Options:

Option Description
-g, --global Remove from global storage and all MCP configs
--mcp Remove MCP server config only (keep skill files)
--for <agents> Remove from specific agents only
-y, --yes Skip confirmation prompt

Note: Use the exact MCP server name as shown in dojo list --mcp (e.g., playwright, not mcp-playwright).


dojo sync

Synchronize skills across all agent formats.

dojo sync              # Sync skills to all detected agents
dojo sync --force      # Overwrite existing skills

Options:

Option Description
-f, --force Overwrite existing skill files

dojo cache <action>

Manage the local registry cache.

dojo cache info        # Show cache statistics
dojo cache clean       # Clear the cache

🔌 MCP Server Integration

Dojo includes its own MCP server for natural language skill discovery. Once configured, agents can respond to prompts like:

"Do you know how to write tests?"
"Teach me debugging"

Auto-Install via CLI

dojo learn --mcp dojo

This automatically configures the MCP server for all detected agents.

Manual Configuration

Claude (~/.claude.json):

{
  "mcpServers": {
    "dojo": {
      "command": "npx",
      "args": ["-y", "@opensourcewtf/dojo-mcp"]
    }
  }
}

Gemini CLI (~/.gemini/settings.json):

{
  "mcpServers": {
    "dojo": {
      "command": "npx",
      "args": ["-y", "@opensourcewtf/dojo-mcp"]
    }
  }
}

📦 Packages

Package Description
@opensourcewtf/dojo CLI for skill management
@opensourcewtf/dojo-mcp MCP server for AI agents

🗃️ Registry

Skills are sourced from the dojo-skills registry, cached via jsDelivr CDN for fast, reliable access.

Registry Categories:

  • official/ — Skills from Anthropic, Google, OpenAI
  • community/ — Community-contributed skills
  • mcp/ — MCP server configurations

🛠️ Development

git clone https://github.com/OpenSourceWTF/dojo.git
cd dojo
pnpm install
pnpm build
pnpm test

Project Structure

packages/
├── cli/     # @opensourcewtf/dojo
└── mcp/     # @opensourcewtf/dojo-mcp

📄 License

MIT © OpenSourceWTF


README copy captured when this server was indexed; it may lag the repository. The original is authoritative.

Questions

About Dojo

How do I install Dojo?

Run claude mcp add dojo -- npx -y @opensourcewtf/dojo-mcp, then add the server to your MCP client's configuration.

Is Dojo safe to use with an AI agent?

Its trust score is 37 out of 100 (low). It passes 1 of the 1 static security checks run so far. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Dojo still maintained?

Conduid hasn't recorded a commit date for this repository yet. Check the repository directly for recent activity.