Packages

adb-tui is a Go-based Android debugging terminal UI (TUI) and an MCP server that exposes broad Android Debug Bridge (ADB) functionality as AI-agent tools (stdio and HTTP/SSE transports). It also provides a scriptable CLI for common ADB operations (devices, shell, screenshot, install, and starting MCP).

Provides a local, read-only Model Context Protocol (MCP) server that exposes Databricks data through CData’s JDBC driver as MCP tools (e.g., get_tables, get_columns, run_query) for LLM clients like Claude Desktop.

DeerFlow (2.0) is an open-source long-horizon “super agent” harness that orchestrates sub-agents, memory, and sandboxed execution to perform complex tasks over minutes to hours, with extensible skills/tools and support for configurable model providers (Python/Node ecosystem) and integrations like MCP servers and messaging (IM) channels.

An MCP server that connects to Garmin Connect (via the python-garminconnect library) and exposes fitness/health data and related operations as a large set of MCP tools for MCP-compatible clients (e.g., Claude Desktop). Includes an auth CLI to pre-authenticate and persist OAuth tokens for MFA and subsequent use.

Hive is a Python runtime harness for AI agents in production. It supports goal-driven agent development (a coding “queen” generates an agent graph/code), then executes that graph with features like state isolation, checkpoint-based crash recovery, cost enforcement/degradation, real-time observability via streaming, and human-in-the-loop pause/intervention nodes. It also advertises integration through MCP tools and tool/agent SDK-wrapped nodes, with support for multiple LLM providers via LiteLLM-compatible interfaces.

Provides a read-only MCP (Model Context Protocol) server that exposes IBM DB2 data to local AI clients via CData JDBC Driver. It runs as a stdio-based MCP server and offers tools to list tables/columns and execute SQL SELECT queries against DB2 without requiring end users to write SQL.

knowledge-mcp is a Python MCP server plus CLI for creating, managing, and querying local knowledge bases backed by LightRAG (hybrid vector + knowledge graph retrieval). Users configure an embedding/LLM provider (e.g., OpenAI), ingest documents into per-KB directories, and query the resulting index via an MCP (FastMCP) interface for use by MCP-capable AI clients.

Provides a read-only MCP (Model Context Protocol) server that exposes LDAP data via a CData JDBC Driver as MCP tools, enabling an MCP-capable client (e.g., Claude Desktop) to discover tables/columns and run SQL SELECT queries against LDAP data (no SQL required in typical usage).

mcp-hub is a local (or self-hosted) management and routing server for Model Context Protocol (MCP). It centralizes lifecycle management of multiple MCP servers (STDIO for local processes and streamable-http/SSE for remote) and exposes a unified MCP endpoint (/mcp) plus a management REST API (/api/*) and event stream (/api/events) for monitoring status and capability changes.

NetBox is an open-source infrastructure resource planning (IPAM/DCIM) web application for managing data center and network assets such as racks, devices, circuits, IP addresses, and related operational metadata.

AIO Sandbox is an all-in-one Docker-based sandbox for AI agents that exposes browser automation (VNC/CDP plus MCP tools), shell execution, file read/write/list/search operations, Jupyter code execution, and an MCP hub. It also provides a web-based VSCode Server and integrates pre-configured MCP servers (browser, file, shell, markitdown) running within the same container with a shared filesystem.

ssh-mcp is a local MCP (Model Context Protocol) server that connects to a remote Linux/Windows host over SSH and exposes tools to run remote shell commands (exec and sudo-exec) from an MCP client.

cdpilot is a CLI for local browser automation using Chrome DevTools Protocol (CDP), offering many browser-control commands (navigation, clicking, typing, screenshots/PDF, console/network/diagnostics, network throttling/proxy, request interception, device emulation, geolocation, cookies/storage, tabs, etc.). It also advertises an MCP server/tool mode so AI agents can invoke the CLI. It launches an isolated browser profile/session locally and communicates with the browser over local CDP endpoints (127.0.0.1 by default).

chatgpt-mcp is an MCP (Model Context Protocol) server for macOS that lets MCP-compatible AI assistants send prompts to the ChatGPT desktop app and retrieve responses. It uses Python/FastMCP and automates interaction with the local desktop app (and notes limited language support).

Provides an “agent skill” (reference documentation files) to help AI agents write, review, and reason about modern Android Jetpack Compose best practices, including state management, effects, navigation, theming, accessibility, and performance.

Kilntainers is an MCP server that provides LLM agents isolated, ephemeral Linux-like execution sandboxes for running shell commands via a single MCP tool (sandbox_exec). It supports multiple backend runtimes including local OCI containers (Docker/Podman), cloud micro-VM sandboxes (Modal/E2B), and WebAssembly-based sandboxes (BusyBox/WASM).

An MCP (Model Context Protocol) bridge that lets an AI client manage file downloads via structured tools: starting multi-thread/concurrent downloads, tracking progress, and managing tasks (list/get/pause/resume/cancel/cleanup), including optional persistent background operation and resumable downloads.

Provides an MCP server for the macOS Fantastical calendar app. It uses AppleScript (and optionally a native EventKit/Swift helper) to create events from natural language, list today/upcoming events, search events, open Fantastical to a date, and list available calendars.

mcp-scanner is a Python tool/SDK/CLI that scans MCP servers and their exposed tools/prompts/resources/instructions for potential security issues. It supports multiple analysis engines (static YARA rules, an LLM-as-judge workflow, and Cisco AI Defense “inspect” API), plus offline JSON scanning, stdio/remote MCP connections, and optional VirusTotal-based malware detection via hash lookups.

Provides an MCP (JSON-RPC 2.0 over stdin/stdout) server that can launch or connect to Microsoft Paint on Windows and control it (e.g., bring window to foreground, draw lines/shapes/pixels, set colors/tools, and query canvas dimensions).