Packages

Provides an MCP server with tools to query public US real-estate related data (housing stats, demographics, nearby amenities, comparisons, cost-of-living/affordability) using US Census ACS and OpenStreetMap (Nominatim/Overpass).

Sift is a reliability gateway that captures JSON outputs from upstream MCP/CLI tools, persists them as SQLite artifacts, redacts secrets on the way back to the model, and returns either inline JSON (`full`) or stable schema-based references (`schema_ref`) for follow-up. It provides explicit pagination continuation until completion and a Python-based artifact querying interface for exact retrieval across one artifact or an entire pagination chain.

A local Model Context Protocol (MCP) server that launches and inspects/controls Tkinter GUI applications. It injects an inspection/controller agent into target Tkinter apps via monkey-patching of tkinter internals, then exposes MCP tools for launching apps, reading the widget tree, capturing screenshots, and interacting with widgets (clicking, focusing, typing, and setting values).

Trading_skills provides an AI-assisted interface (Claude Code/Cursor integration and a Claude Desktop MCP server) for market and options analysis. It pulls market data (Yahoo Finance) for quotes/history/fundamentals/news/earnings and computes technical indicators, option Greeks, spread analysis, and risk metrics; it can also connect to Interactive Brokers via TWS/IB Gateway to summarize a portfolio and generate event/earnings-aware action reports and option roll candidates (per README descriptions).

An MCP server that exposes Oracle Cloud Infrastructure (OCI) Object Storage service APIs, enabling AI agents to create and manage buckets, upload and download objects, configure lifecycle policies, manage access policies, set up pre-authenticated requests, and interact with OCI's S3-compatible object storage at scale.

AgentRPC is a universal RPC layer that allows AI agents to connect to functions deployed anywhere (private VPCs, Kubernetes, multi-cloud) across network boundaries and language barriers. It wraps functions in an RPC interface exposed via MCP and OpenAI-compatible tool definitions.

Config-driven backtesting framework for algorithmic trading built on Backtrader, with an integrated MCP server that lets LLMs run and analyze backtests across stocks, crypto, and forex.

Alph (@aqualia/alph-cli) is a local-first CLI that configures Model Context Protocol (MCP) servers for multiple AI coding agents by detecting agents, validating configuration, performing atomic file writes with timestamped backups, and supporting rollback. It supports MCP transports (HTTP, SSE, and local STDIO) and can run in interactive or non-interactive modes, with a dry-run option to preview changes.

Provides an MCP server (usable via npx) that lets an MCP client query locally exported Apple Health data from CSV files using natural-language-to-SQL plus direct SQL execution, with schema introspection and automated health reporting. Uses DuckDB for in-process analytics, with lazy loading by time windows and result caching via TTL.

A drop-in replacement/upgrade for the official Claude Code Telegram plugin that adds richer Telegram behaviors (threading, batching, media handling, reactions/status, history/search/memory, scheduling, daemon management) plus voice input/output (transcription via OpenAI Whisper or local whisper tools; TTS via ElevenLabs). It is integrated as an MCP-style “channel”/plugin for Claude Code and provides tools to the assistant such as reply, react, edit_message, ask_user, get_history, search_messages, clear_history, save_memory, and create_telegraph_page.

FreeCAD-MCP is a FreeCAD workbench/plugin that exposes FreeCAD automation capabilities through the Model Context Protocol (MCP). It includes an MCP server with a GUI control panel, plus a Python client that can create/update/validate/run FreeCAD macros (FCMacro), adjust 3D views, and retrieve execution logs/reports. Communication is via stdio or TCP and integrates macro normalization (imports, recompute, view adjustment) before running.

Provides a Spring Boot Model Context Protocol (MCP) server (streamable HTTP transport) exposing Greenplum/PostgreSQL database tools (schema discovery and query execution) with API-key authentication, policy enforcement (query validation, row/byte limits, allow-lists, redaction), and observability features (health, metrics, tracing).

kube-apiserver is the Kubernetes API server component that exposes the Kubernetes control-plane API (REST over HTTPS) used by kubectl, controllers, and other clients to manage cluster resources. It implements authentication, authorization, admission, persistence via etcd, and core admission/validation pathways.

mcp-anything is a CLI/code generator that wraps an existing codebase or API specification into a self-contained MCP server (FastMCP/HTTP SSE or mcp-use/HTTP). It can generate servers from local frameworks/CLIs (FastAPI, Flask, etc.) or from API specs (OpenAPI/Swagger, GraphQL SDL, gRPC/protobuf), and includes tooling for scoping exposed capabilities, editing tool descriptions, and producing an AGENTS.md for tool discovery.

A NestJS module for building and exposing Model Context Protocol (MCP) servers that map NestJS methods to MCP tools (with Zod validation), plus MCP resources and prompt templates. It supports multiple transports (HTTP+SSE, Streamable HTTP, STDIO) and offers guard-based and built-in/external authorization patterns (including OAuth).

Provides an MCP server for querying the Open Food Facts product dataset. It downloads/caches a Parquet product database and uses DuckDB for fast search, exposing MCP tools for brand/name lookup and barcode lookup. Supports local STDIO (Claude Desktop) mode without auth and remote HTTP mode with Bearer-token auth (except /health).

outlook-mcp is an MCP server that lets AI assistants interact with Microsoft Outlook email and calendar (and optionally SharePoint files) via the Microsoft Graph API. It supports reading/searching/sending emails, managing calendar events, downloading/parsing attachments and SharePoint files, and saving large downloads to a local work directory to avoid MCP response size limits.

Provides an MCP (Model Context Protocol) server exposing Public.com brokerage/trading API operations as MCP tools (read-only and write/destructive actions) over stdio for AI assistants.

An MCP (Model Context Protocol) server that exposes tools for querying Terraform Registry information—provider documentation and module discovery/details—so an AI agent can retrieve and reason over Terraform Registry metadata for IaC development and automation.

Provides an MCP (Model Context Protocol) server that integrates with TickTick (and optionally Dida365) via TickTick’s/Open API, exposing MCP tools to view, create, update, complete, and delete projects/tasks.