Personal AI coding agent runtime: IDE-grade code editing, isolated Git worktree subagents, time-traveling stream rules, and a dual-direction OpenAPI and OpenClaude gateway.
Drive an MCP server over real stdio against a real backend, and prove every tool was exercised
Playwright MCP-compatible browser automation bridge for CloakBrowser Chromium.
Sub-Sub: a research assistant that works from your own Zotero library, in the browser or the terminal, built on pi
Signed, offline-verifiable authorization for API-contract changes. Only a granted change can proceed. This package is the TypeScript client that preflights a change set and verifies its receipt.
The harness package manager: share agent rules, docs, skills and tools through npm packages
Headless, configurable in-browser LLM agent on the Vercel AI SDK: multi-provider bring-your-own-key (OpenAI, Anthropic, Google, xAI, DeepSeek, gateway, openai-compatible) AND local WebGPU/WebLLM, nati
ai-hub — work in your AI Hub Second Brain from the terminal, with the same verbs your AI agents use.
TypeScript SDK for Ratel — context engineering platform for AI agents. BM25 tool retrieval, MCP ingestion, framework-neutral capability tools.
Runs CodeZaiku, a coding and development harness on local models (Ollama, llama.cpp) or hosted APIs, as an MCP server, so Claude Code, Codex or Gemini CLI can hand it coding, review and ops tasks.
OS keychain secrets for AI coding agents, over MCP — Cursor, Kiro, Claude Code. Native vault storage, policy-governed access, audit log.
An on-device agent for the browser: a tool-calling harness over WebMCP and remote MCP, plus a drop-in chat UI. Decides whether to run a tool, ask for what's missing, or confirm before writing.
MCP server that gives Claude Code and other MCP hosts long-context memory of a codebase through Gemini's 1M-token window: `ask`/`code` send the scanned workspace with implicit caching by default or an
A receipt for every AI agent action. One line: every tool call gets a sealed, publicly checkable receipt. Vercel AI SDK, OpenTelemetry, LangChain.js, Mastra, MCP.
The decision proxy behind VERAX: policy, signed decision records, and the ledger they live in.
Manage MCP servers and skills in one place; hide or restore groups of tool definitions to control what enters the next prompt.
Open-source CLI, MCP, skills, templates, workflows, recipes, and evals layer for giving AI agents reliable access to protected web data through Zenrows cloud infrastructure.
Expose Elysia HTTP routes, prompts, and resources through a native-feeling MCP adapter.
A local-first trust, provenance, and review gate for AI agent tools, MCP servers, skills, and plugins.
When the agent writes a bad import, the write doesn’t land. The same check fails the pull request.
CTM + Wall-E: one workspace for coding agents, worktrees and review, with shared session context and a personal agent.
Code-mode MCP server for the U.S. Federal Register, Electronic Code of Federal Regulations, and regulations.gov APIs.
MCP server that gives AI agents full access to the Fastly API
BuiltWith AI-first SDK for Node.js
Local-first code retrieval and coding agent — hybrid semantic + keyword search over your codebase, served to any AI over MCP. Runs fully offline.
Command line interface for PgBeam, a Postgres proxy for safe AI agent access with policy enforcement in the wire protocol
Agent runtime for Gaunt Sloth: tools, middleware, MCP/A2A clients and the AG-UI server
An MCP server for OPF: learn the format (docs, schemas, the default catalog, examples and agent skills as resources and tools), validate and convert decks. Runs locally on stdio (opf-mcp) or hosted on
Real-time context window breakdown, prompt & thinking effort advisor, MCP server controller, and 5h/weekly quota telemetry sidecar for Google Antigravity.
Cross-platform clipboard history tray app with numpad quick-paste and native cloud sync
MCP(Model Context Protocol)客户端适配器:基于官方 @modelcontextprotocol/sdk 实现连接;McpManager(mm_machine 管理器)读取 mcp.json 连接配置,编排多台外部 MCP 服务器,支持 stdio / Streamable HTTP / HTTP+SSE,聚合工具、资源、提示词,配置即接入并支持热更新,可直接接入大模型做工
Restores reasoning content dropped from multi-turn OpenCode conversations (DeepSeek, Kimi, GLM, MiMo, MiniMax, InclusionAI/Ling, OpenCode Go, OpenRouter). V1 + V2 plugin, proxy, and watchdog.
Experimental: MCP interceptor server (SEP-2624 draft) that labels tool results and checks tool calls with ADTL trust labels
ActionDock SDK for defining and testing standalone AI Agent Actions
Model Context Protocol server for managing a todo.txt file
米家自动化极客版 MCP Server:在任意 AI Agent 里用说话的方式创建和修改米家中枢网关的自动化规则
Direktori lengkap
IDE, git, linting, dan analisis kode. 8,613 server, yang paling banyak dipakai lebih dulu. Masing-masing membuka halaman penerbit.