Browse MCP servers — page 78

An MCP Server that enables AI assistants to interact with your local browsers. Note: Consider using a separate browser profile or dedicated browser instance with this MCP to prevent sensitive data from being unintentionally exposed to AI model providers.

MCP to use the python debugger, inspired by debug-gym. This tool executes Python code through the debugger and should be used in trusted environments only.

用于检查多个顶级域（TLD）中域名可用性的MCP服务器。需要Python 3.10+版本和uv包管理器进行依赖管理。支持通过DNS和WHOIS验证，检测50多个常用TLD的域名可用性。

oh-my-dag — a DAG execution engine exposed as an MCP server. Typed nodes → concurrent fan-out → cross-model verify → escalate on failure; pathfinder mode (persistent decision maps for multi-session work), an agent-template registry, a memory layer, and a bundle of /omd-* client skills that install themselves into your MCP client. Any OpenAI-compatible backend plugs in, plus OAuth providers via the pi-ai catalog. Front ends: any MCP client, or omd's own TUI (omd tui).

Okta MCP 服务器使 AI 模型能够通过模型上下文协议（MCP）直接与您的 Okta 环境交互。该插件专为身份与访问管理（IAM）工程师、安全团队和 Okta 管理员设计，用于管理和分析 Okta 资源。需要在您的机器上安装 Python 3.8 及以上版本。配置环境变量以获取 Okta 租户 API 访问权限和 AI 提供商凭据。

OpenCTI MCP Server 是一个模型上下文协议（Model Context Protocol）服务器，提供与 OpenCTI 平台的集成，用于查询和检索威胁情报数据。需要通过环境变量 OPENCTI_URL 和 OPENCTI_TOKEN 进行外部配置，以访问 OpenCTI 实例。

A Model Context Protocol (MCP) server implemented in Go, providing tools to analyze Go pprof performance profiles including CPU, heap, goroutine, allocs, and more. Requires Graphviz installed for flamegraph generation (dot command must be in PATH).

PyMOL-MCP 将 PyMOL 连接到 Claude AI，通过模型上下文协议（MCP）实现对话式结构生物学、分子可视化和分析，支持通过自然语言操作。需要在系统中安装 PyMOL 以及在桌面版 Claude 中配置服务器脚本路径。

用于 cosense/scrapbox 的 MCP 服务器，提供页面内容检索、列表、搜索、创建和 URL 生成功能。需要环境变量进行项目配置和身份验证，包括用于私有项目的 COSENSE_PROJECT_NAME 和 COSENSE_SID。

一个 Model Context Protocol (MCP) 服务器，连接至 Tana 的 Input API，允许大型语言模型（LLMs）和其他 MCP 客户端在 Tana 工作区中创建和操作数据。需要通过环境变量 TANA_API_TOKEN 配置的 Tana API 令牌。

Thoughtbox is an MCP server providing structured reasoning tools for AI agents, enabling step-by-step thinking, branching explorations, revisions, and autonomous critique. It runs locally, storing data at ~/.thoughtbox/ by default, ensuring privacy and control.

Trustabl is an open source AI agent reliability tool that scans agent repositories for guardrail gaps, unsafe tools, missing validation, and unbounded loops, then automatically fixes them. It runs fully locally with no cloud or LLM required.