MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
MonkeyPlannergithub.com/kjm99d/MonkeyPlannerupdated this year
Local-first task manager for AI coding agents. Native MCP server for Claude Code, Cursor, and Continue.
go MIT ★ 1 -
CURSOR25X is an interactive task loop MCP (Model Context Protocol) server developed for Cursor IDE. It performs task-based operations by continuously receiving user input and develops modern web applications.
MIT ★ 21 -
Zenith: a continuous-improvement harness for long-running agent tasks. Turns Claude Code, Codex, or Hermes into a multi-agent mission orchestrator via MCP/ACP.
Apache-2.0 ★ 403 -
MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI (agy) — purpose-built tools, model routing with fallback, and multi-turn session continuity.
npm MIT -
A Codex plugin and local MCP server for deliberate context compaction in VS Code, with a verified task handoff and one-time automatic continuation.
MIT ★ 1 -
MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI (agy) — purpose-built tools, model routing with fallback, and multi-turn session continuity.
npm MIT -
Local MCP for Chief of Staff agents to create, run, monitor and continue Codex projects and tasks.
npm MIT -
Local-first task continuity and reviewed project knowledge for MCP coding agents. 织忆:让 AI 接上上次的任务,也接上项目经验。
AGPL-3.0 -
Local workspaces for AI agents. Preserve file versions, continue tasks across conversations, rebuild affected steps, and reuse workflows through CLI and MCP.
MIT ★ 1 -
Local-first MCP server that lets one AI coding agent (Claude, Copilot, Cursor, ...) continue another agent's task from its last checkpoint
npm MIT -
Provide a simple Python library to handle tasks efficiently with full type support, continuous testing, and complete code coverage.
-
Unofficial Windows stdio MCP bridge for delegating Codex tasks to local Unsloth Qwen models with aliases, safe progress, cancellation cleanup, bounded continuation, and single-flight execution.
AGPL-3.0 -
MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI (agy) — purpose-built tools, model routing with fallback, and multi-turn session continuity.
npm MIT ★ 8 -
One self-hosted memory for every AI app (Claude, ChatGPT, Gemini, Cursor, Codex, Copilot) over MCP, with tasks any app can continue when another hits its limit.
AGPL-3.0 ★ 5 -
Local-first AI memory system for coding agents: desktop app, CLI, and MCP server for durable project context, handoffs, recall, and task continuity.
MIT ★ 11 -
MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI (agy) — purpose-built tools, model routing with fallback, session continuity, and output truncation to save Claude's context and tokens.
npm MIT ★ 67 -
MCP providing a capability layer between agents and their environment. Giving governed access to tools, scripts, skills, execution, Runs and Task continuity without coupling the product to one agent or one MCP gateway.
MIT -
gc-mcpgithub.com/Genesis-Conductor-Engine/gc-mcpupdated this year
Universal AI MCP server for Genesis Conductor — 7 gc_* tools (task submission, Soul Capsule WPP/EU-AI-Act traces, agent routing, telemetry). stdio + streamable HTTP. Claude Desktop / Cursor / Continue native. Baremetal-validated on diamondnode GTX 1650.
NOASSERTION -
Local task-state ledger for AI coding agents.
npm MIT -
A CLI-first autonomous coding agent where the actor never grades itself. Every edit is grounded in your repo's real topology and verified by an independent deterministic grader (drift, lint, verify-task, intent) before it ships. Closed-loop: comprehend → plan → execute → critique → replan. Also a passive MCP harness.
go Apache-2.0 ★ 25