"Scratch" matching MCP servers:

Enables clients to interact with a bundled documentation corpus through MCP tools and resources over HTTP or stdio, and provides a trajectory evaluation harness for grading agent performance.

A fully working MCP server built from scratch in plain Node.js, implementing tools, resources, prompts, notifications, and sampling according to the MCP specification, designed to connect to Claude Desktop or any MCP client.

Enables learning and integration testing of the Model Context Protocol by evolving from a minimal server toward production-grade tools, resources, prompts, and agent interoperability.

A Model Context Protocol server that provides a scratchpad tool for LLMs to organize thoughts, take notes, and plan approaches to complex problems.

Enables AI assistants to scout GitHub for existing open-source projects, libraries, and frameworks before building from scratch, offering tools for stack suggestions, repository searches, and detailed repo evaluation.

A Python MCP server implemented from scratch, providing tool execution capabilities like system time and file hashing via stdio or HTTP transport.

Provides AI coding agents with structured, deterministic repository intelligence—file tree, dependencies, hotspots, and git history—via Model Context Protocol, enabling them to understand codebases without reconstructing architecture from scratch.

A simple MCP server that exposes stock price lookup tools, demonstrating how to set up MCP from scratch with Ollama.

Enables shared scratchpad functionality for Claude Code sub-agents with full-text search, workflow organization, and intelligent Chinese text tokenization. Supports seamless context sharing and collaboration across workflows with project scope isolation.

Manages a global scratchpad file on your Desktop for tracking interruptions, ideas, tasks, and focus across all projects with secure input validation and rate limiting.

Connects to Salesforce scratch orgs and allows AI to generate and insert realistic nested test records via natural language descriptions using the Composite Tree REST API.

An educational MCP server that teaches implementing the Model Context Protocol from scratch, enabling AI models to discover and invoke tools via JSON-RPC over stdio.

Enables structured input surfaces inside Claude conversations with prefilled forms, assumption ledgers, and review matrices, so the agent can infer values and present proposals for auditing rather than authoring from scratch.

An MCP server that exposes the markdown issue tracker under .scratch/ as a queryable graph, letting agents see open, blocked, or takeable issues without re-parsing prose.

Lets users discover, compare, and audit public MCP servers from the official registry, validate server.json metadata, and generate installation snippets and publication checklists.

Enables AES-256-CBC encryption and decryption of messages with passphrase-derived keys, exposing tools, resources, and prompts for MCP clients like VS Code Copilot Chat.

Generates crash-proof Python LLM pipelines from plain English descriptions. Pipelines resume from the last successful node after a crash instead of restarting from scratch. IProvider-agnostic , works with Groq, Gemini, Ollama, OpenAI, or any Python callable. Zero infrastructure, runs entirely on free-tier APIs.

A demonstration MCP server providing basic tools for returning greetings and performing mathematical addition. It serves as a practical workshop guide for setting up and configuring Model Context Protocol servers from scratch using Python.

Enables coding agents to search, inspect, and retrieve real UI components from existing component libraries, with ranked results, score explanations, licence and provenance details, and safe installation plans. It also surfaces similar components and higher-level UI patterns so agents can reuse existing interface code instead of generating it from scratch.

Gives Claude Code sessions a durable, shared memory and work ledger backed by a local SQLite database, exposing the brain_* tools so agents can create and lock goals with verifiable success criteria, plan and track work units, record evidence and failures, and search decisions, learnings and project knowledge. Because the tool descriptions carry the operating rules, every session in every project reads and writes the same accumulating context instead of starting from scratch.