Packages

**A common shape for agents that live alongside people.**

Provides a local MCP (Model Context Protocol) server that exposes DaisyUI component documentation as MCP tools. It can fetch DaisyUI’s public `llms.txt` and store component docs as local Markdown, then serve two tools: list components and get component documentation.

FastMCP is a Python framework for building MCP (Model Context Protocol) servers (exposing tools/resources/prompts) and connecting to MCP servers with clients. It focuses on turning Python functions into MCP tools with automatic schema/validation/documentation generation, and handling MCP transport/authentication/lifecycle details for you.

Provides an experimental MCP (Model Context Protocol) stdio server that exposes Granola.ai meeting intelligence stored in local cache files (search, metadata, transcripts, documents/notes, and cross-meeting pattern analysis) to MCP-capable clients like Claude Desktop.

Provides an MCP (Model Context Protocol) server for Vuetify that exposes tools to query Vuetify component/directive/API documentation, installation guides, FAQs, and release notes to MCP-compatible clients/IDEs. Supports local stdio (default) and HTTP transports (hosted or self-hosted) and optionally uses a Vuetify API key for certain tools.

[](https://github.com/ubgb/moltmemory/stargazers)

A self-hosted web UI for managing Nginx servers with real-time monitoring, online configuration editing with validation, SSL certificate automation via Let's Encrypt, multi-server cluster support, and an MCP interface for AI-assisted configuration management.

MCP server for Elasticsearch 7.x, enabling AI agents to perform search queries, index documents, manage indices, and interact with Elasticsearch clusters running version 7. Provides a natural language-friendly interface for search and document operations against ES7 clusters.

The free and open-source React Best Practices Agent built on 10+ years of real-world experience distilled in Skills powered by Bluebag

Secure OpenClaw gateway - WhatsApp, Telegram, Signal, iMessage adapters with Claude agent runtime

xenodocs-mcp-server is an MCP server that lets an AI agent query a documentation database for up-to-date, version-specific documentation and code examples for Python (and potentially other) libraries. It exposes MCP tools to search for libraries by name and to search within a library’s docs for a query.

This package provides a Python-based MCP server for “Xiaoya” (小雅) education management, exposing tools for teacher-oriented workflows such as course resource management, question/quiz creation and management, classroom & attendance (签到) queries, and task/assignment + grading workflows. It supports multiple MCP transports (stdio by default, plus SSE and streamable HTTP). Authentication to the upstream service can be done either via account/password (with server-side token caching and automatic re-login) or via a provided Bearer token.

Provides an MCP (Model Context Protocol) server for YNAB that exposes the YNAB API as MCP tools, auto-generating tools from YNAB’s OpenAPI spec via FastMCP. It uses a YNAB personal access token (env var) to authenticate requests and includes support for common CRUD operations across users, budgets, accounts, categories, transactions, payees, scheduled transactions, and budget months.

Agent Board is a multi-agent task orchestration service combining a Kanban board with DAG-enforced task dependencies, quality gates, auto-retries, task chaining, signed webhooks for event-driven wakeups, and an append-only audit trail. It exposes both a REST API and an MCP server so AI agents can create/move tasks and manage task comment threads.

Desktop application that extends Claude with persistent background automation, scheduled task execution, memory management, file watching, and worker agents. Runs as a menu bar/system tray app with a Node.js sidecar for scheduling and an MCP server for Claude integration. Supports pre-configured worker personalities (researcher, code reviewer, email assistant, etc.) and self-improving tasks that optimize based on execution history.

Provides an MCP server on macOS that lets AI assistants access Apple Mail via natural-language tools: read/search emails, manage folders and message state, compose/reply/forward, handle drafts and attachments, and generate basic analytics/export.

Provides an MCP server that lets an AI assistant query BloodHound Community Edition data (via BloodHound CE REST API and Neo4j graph) through a set of composite MCP tools (e.g., domain/user/group/computer info, graph analysis, ADCS info, and Cypher execution) and reference resources for Cypher and AD/ADCS/Azure methodology.

A Python demo/template repo for configuring and “mastering” Anthropic Claude Code hooks. It provides a full hook lifecycle implementation (13 hook events), logging, transcript conversion, prompt validation/context injection, tool permission auditing/auto-allow for read-only ops, and examples of orchestration via sub-agents and team-based validation. It also demonstrates optional TTS via multiple providers (e.g., ElevenLabs) using MCP servers.

Provides an MCP server that exposes two Codex-related tools to MCP-compatible coding assistants, bridging tool calls to the official OpenAI Codex CLI via subprocess execution. Supports configurable timeouts and (optionally) bypassing Codex CLI git-repository checks.