1. Conduid
  2. AI
  3. Factspark MCP Toolkit
MCP server Β· AI

Factspark MCP Toolkit

πŸ”§ Swiss Army Knife for MCP Servers β€” Test, Debug, Scaffold & Discover Model Context Protocol servers

37Low

Scored 5 months ago Β· breakdown

About Factspark MCP Toolkit

Factspark MCP Toolkit is an MCP server in the AI category: πŸ”§ Swiss Army Knife for MCP Servers β€” Test, Debug, Scaffold & Discover Model Context Protocol servers.

Install

Claude Code
claude mcp add factspark-mcp-toolkit -- npx -y factspark-mcp-toolkit
npx
npx -y factspark-mcp-toolkit

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 Factspark MCP Toolkit

Powered by Claude · Grounded in docs

I know everything about Factspark MCP Toolkit. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

  • βœ“README presentThe repository ships documentation.

5 checks have not been run yet.

README

πŸ”§ MCP Toolkit

The Swiss Army Knife for Model Context Protocol Servers

Test Β· Inspect Β· Scaffold Β· Discover Β· Monitor

npm version Node.js GitHub stars GitHub issues

Quick Start Β· Commands Β· Why MCP Toolkit? Β· Contributing


Building MCP servers? You know the pain:

  • No way to test your server without writing a full client
  • No standard tool to validate protocol compliance
  • Boilerplate hell every time you start a new server
  • Finding servers means digging through GitHub repos

MCP Toolkit fixes all of that. One CLI, five commands, zero friction.

⚑ Quick Start

# Install globally
npm install -g factspark-mcp-toolkit

# Test any MCP server
mcp-toolkit test npx @modelcontextprotocol/server-filesystem /tmp

# Interactive inspector
mcp-toolkit inspect npx @modelcontextprotocol/server-filesystem /tmp

# Scaffold a new server
mcp-toolkit scaffold my-server --language typescript

# Browse the registry
mcp-toolkit discover --category database

# Health check
mcp-toolkit health npx @modelcontextprotocol/server-memory

🎯 Commands

mcp-toolkit test β€” Validate Everything

Runs a full compliance check on any MCP server:

mcp-toolkit test ./my-server.js
mcp-toolkit test npx @modelcontextprotocol/server-github
mcp-toolkit test python server.py
mcp-toolkit test http://localhost:3000

What it checks:

  • βœ… Server starts and completes handshake
  • βœ… Protocol version compatibility
  • βœ… All tools have valid schemas
  • βœ… Tools respond to calls
  • βœ… Resources are readable
  • βœ… Prompts return messages
  • βœ… Response time benchmarks

Output:

Test Results
────────────────────────────────────────────────────────
  βœ“ connection
  βœ“ server-info          my-server v1.0.0
  βœ“ capabilities         tools=3 resources=1 prompts=2
  βœ“ tool:greet           Responded to empty call in 45ms
  βœ“ tool:calculate       Schema validates required params
  βœ“ resource:info://app  Readable β€” 1 content block(s)
  βœ“ prompt:summarize     Returns 1 message(s)

────────────────────────────────────────────────────────
  7 passed

JSON output for CI:

mcp-toolkit test ./server.js --json | jq '.[] | select(.status == "fail")'

mcp-toolkit inspect β€” Interactive Debugger

Browse and call tools, read resources, test prompts β€” all interactively:

mcp-toolkit inspect npx @modelcontextprotocol/server-filesystem /tmp
πŸ” Interactive MCP Inspector

βœ“ Connected β€” 5 tools, 1 resources, 0 prompts

? What do you want to do? (Use arrow keys)
  πŸ”§ Browse Tools (5)
  πŸ“¦ Browse Resources (1)
  πŸ’¬ Browse Prompts (0)
  πŸ“Š Server Info
  ─────────────
  πŸš” Exit

mcp-toolkit scaffold β€” Zero to Server in Seconds

Generate production-ready MCP server projects:

# TypeScript minimal
mcp-toolkit scaffold my-server --language typescript --template minimal

# Python full (with prompts, tests, etc.)
mcp-toolkit scaffold my-server --language python --template full

# Skip npm install
mcp-toolkit scaffold my-server --no-install

Templates:

Language Template What's Included
TypeScript minimal Server, tool, resource, README
TypeScript full + prompts, tests, eslint, vitest
Python minimal Server, tool, resource, README
Python full + prompts, tests, pytest

mcp-toolkit discover β€” Server Registry

Find MCP servers from a curated, tested registry:

# Browse all
mcp-toolkit discover

# Search
mcp-toolkit discover --query filesystem

# Filter by category
mcp-toolkit discover --category database

# JSON output
mcp-toolkit discover --json

Categories: filesystem Β· database Β· devtools Β· search Β· browser Β· api Β· ai Β· communication Β· web Β· reference

mcp-toolkit health β€” Monitor Servers

# One-time check
mcp-toolkit health ./my-server.js

# Continuous monitoring
mcp-toolkit health ./my-server.js --watch --interval 10
── Health Report ──
  Status:    ● HEALTHY
  Latency:   142ms
  Server:    my-server v1.0.0
  Tools:     3
  Resources: 1
  Prompts:   2
  Checked:   2026-04-23T15:30:00.000Z

πŸ€” Why MCP Toolkit?

Feature MCP Toolkit Manual Testing Other Tools
Protocol compliance testing βœ… Built-in ❌ Write your own ❌ None
Interactive inspector βœ… REPL UI ❌ Console.log ⚠️ Web only
Server scaffolding βœ… 4 templates ❌ Copy-paste ⚠️ Limited
Server registry βœ… Curated ❌ Search GitHub ⚠️ Lists only
Health monitoring βœ… Watch mode ❌ Manual ❌ None
CI/CD ready βœ… JSON output ❌ No ⚠️ Partial

πŸ”Œ Works With Everything

# Node.js servers
mcp-toolkit test ./dist/index.js

# Python servers
mcp-toolkit test python server.py

# npm packages
mcp-toolkit test npx @modelcontextprotocol/server-github

# Remote servers (SSE)
mcp-toolkit test http://localhost:3000

# Custom commands
mcp-toolkit test "docker run -i my-mcp-server"

πŸ—οΈ For CI/CD

Use --json flag for machine-readable output:

# In your CI pipeline
mcp-toolkit test ./server.js --json > test-results.json

# Fail build on any test failure
mcp-toolkit test ./server.js && echo "PASS" || echo "FAIL"
# GitHub Actions example
- name: Test MCP Server
  run: |
    npm install -g factspark-mcp-toolkit
    mcp-toolkit test ./dist/index.js

πŸ“– Programmatic API

import { connect, getServerInfo } from 'mcp-toolkit';

const conn = await connect('./my-server.js');
const info = await getServerInfo(conn.client);

console.log(`Tools: ${info.tools.length}`);
console.log(`Resources: ${info.resources.length}`);

// Call a tool
const result = await conn.client.callTool({
  name: 'greet',
  arguments: { name: 'World' },
});

await conn.close();

🀝 Contributing

We love contributions! Here's how:

  1. Add a server to the registry β€” Edit src/commands/discover.ts
  2. Add a scaffold template β€” Edit src/commands/scaffold.ts
  3. Improve test coverage β€” Add tests in tests/
  4. Report bugs β€” Open an issue
# Development setup
git clone https://github.com/factspark23-hash/mcp-toolkit.git
cd mcp-toolkit
npm install
npm run dev

See CONTRIBUTING.md for details.

πŸ“‹ Roadmap

  • Web UI for interactive inspection
  • VS Code extension
  • Server performance benchmarks
  • Auto-generate client code from server
  • Docker image for CI/CD
  • More server templates (Rust, Go, Java)
  • Server compatibility matrix

⭐ Star History

If MCP Toolkit helps you build better MCP servers, give it a star! ⭐

πŸ“„ License

MIT β€” use it however you want.


Built for the MCP ecosystem

Model Context Protocol Β· MCP Servers Β· MCP Spec

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

Questions

About Factspark MCP Toolkit

How do I install Factspark MCP Toolkit?

Run claude mcp add factspark-mcp-toolkit -- npx -y factspark-mcp-toolkit, then add the server to your MCP client's configuration.

Is Factspark MCP Toolkit 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 Factspark MCP Toolkit still maintained?

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