MCP Tools | Model Context Protocol Server Directory

The MCP Notion server bridges the gap between your LLM context and your structured knowledge base. Instead of manually copying and pasting documentation or project requirements, this tool allows your AI agent to programmatically query, read, and update Notion pages and databases directly. For developers, this means your local development environment can interact with your team's existing project wikis, sprint backlogs, or technical specs in real-time. Unlike standard API integrations that require custom boilerplate for every agentic workflow, this MCP implementation provides a standardized interface that works out-of-the-box with any MCP-compliant host. It transforms Notion from a passive documentation silo into an active, searchable memory layer for your AI-driven development processes, enabling more coherent task management and automated knowledge retrieval.

The MCP Jira tool bridges the gap between your local LLM environment and your project management workflow. Instead of context-switching between your IDE and a web browser to check task statuses, this protocol implementation allows your AI assistant to interact directly with your Jira instance. Developers can use it to query issue details, transition ticket statuses, create new sub-tasks, or search for specific epic requirements using natural language commands. Unlike standard API wrappers that require manual script writing, this MCP implementation provides a standardized interface that lets your model understand the structure of your Jira projects. It is particularly effective for automating repetitive administrative tasks—like updating progress during a coding session—or using project context to ground the AI's suggestions in your actual sprint goals. Integration is straightforward via any MCP-compliant host, making it a powerful addition to a developer-centric agentic workflow.

FlashDesk introduces a critical bridge between LLM-driven coding and remote GUI environments. While SSH is the standard for headless terminal operations, it falls short when your workflow requires interacting with visual applications or complex desktop interfaces. FlashDesk solves this by implementing the Model Context Protocol to allow Claude Code to perform direct desktop manipulation—including screenshots, mouse movements, keyboard input, and file transfers—on a remote machine. Unlike standard remote desktop tools that require manual human intervention, this MCP server allows the AI to 'see' and act upon a remote workspace as if it were local. Security is handled via a per-connection approval flow on the target machine, ensuring the LLM cannot hijack a system without explicit authorization. For developers managing remote build servers, testing UI components, or automating legacy desktop software via natural language, FlashDesk transforms Claude from a code generator into a functional remote operator.

For developers working with LLMs on Windows, the biggest hurdle is often the trade-off between agentic autonomy and system security. DeskMCP addresses this by acting as a controlled policy gateway for the Model Context Protocol. Unlike standard implementations that might grant an LLM broad access to your filesystem, DeskMCP enforces workspace-scoped permissions and guarded write operations. This means you can safely allow ChatGPT or other MCP-compatible clients to interact with specific local files and terminal sessions without risking your entire environment. It manages session-owned processes, ensuring that any command executed by the model is tied to a specific context and tracked via audit logs. Whether you are automating complex build scripts or managing local documentation, DeskMCP provides the necessary sandbox to let AI tools work locally while maintaining strict oversight of what can be read, written, or executed.

This MCP server wraps QoreNext's trade-screening API so you can run real-time sanctions and due-diligence checks against entities, persons, and transactions without leaving your chat context. It exposes a small set of tools—like screening a company name, fetching detailed hit profiles, or validating transaction parties—that any MCP-aware client (Claude Desktop, custom agents, or internal bots) can call through the standard protocol. Use it when you're prototyping compliance workflows, embedding screening into conversational UIs, or building agents that need to verify counterparties on the fly. Compared to hitting the REST endpoint manually, the MCP integration removes boilerplate auth handling and response parsing, keeps state in-chat, and lets you chain screening steps with other MCP-backed tools (retrieval, summarization, etc.). It's stateless per request, returns structured JSON, and handles fuzzy matching plus alias resolution out of the box. You bring your own API key, and it's lightweight enough to drop into existing agent runners without extra infra.

One is a community-built MCP server that lets your agent talk to over 700 SaaS apps—think Gmail, Slack, Stripe, Shopify, HubSpot, Notion, and Linear—through a single integration point. Instead of wiring up individual API clients per service, you authenticate once via OAuth and gain access to four composable tools that cover the majority of common workflows: reading and writing records, triggering actions, listing resources, and fetching metadata. It slots into any MCP-compatible runtime (Claude Code, Cursor, custom agents) and handles token refresh, pagination, and error normalization behind the scenes. For developers juggling multi-app automations, it cuts boilerplate significantly compared to managing each API separately. The tradeoff is less granular control than first-party SDKs, but for prototyping, internal tools, or agentic workflows where speed matters more than fine-tuned performance, it's a pragmatic shortcut. Since it's community-maintained, expect occasional gaps in coverage and faster iteration than enterprise connectors.

The mcp-n8n MCP server exposes 55 tools that let you build, edit, and run n8n workflows directly from your assistant. It handles node validation, surgical edits to existing workflows, webhook testing, and rollback of changes. You can create and modify workflows, manage credentials, trigger executions, and inspect run logs without leaving your development environment. It integrates with the n8n REST API, so you configure connection details once and then operate n8n programmatically through MCP-compatible clients. Unlike generic API wrappers, this server provides workflow-aware operations like node-level edits and automatic validation, reducing manual JSON editing. It's useful for automating deployment pipelines, testing integrations, and rapidly prototyping automation flows. The server is community-maintained and requires a running n8n instance with API access enabled.

Auditra exposes a focused MCP interface for auditing WordPress plugins directly from your IDE or AI assistant. Instead of manually scanning codebases or running disparate CLI tools, you get a standardized `audit_plugin` method that accepts a plugin slug or local path and returns structured findings: outdated dependencies, known CVEs, coding-standard violations, deprecated API usage, and permission overreach. The server wraps WP-CLI, PHP_CodeSniffer (WordPress ruleset), Composer audit, and a curated vulnerability database, so results are consistent and actionable. Typical workflows include pre-merge checks in CI, quick vetting of third-party plugins before installation, and automated reporting for client site maintenance. Integration is zero-config for most stacks — add the server to your MCP client (Cursor, Claude Desktop, Continue, etc.) and call the tool like any other function. Compared to standalone SAST scanners, Auditra is WordPress-aware: it understands plugin architecture, hook semantics, and the WP.org update cycle, reducing false positives. It doesn't replace deep static analysis but excels at the 80% of issues that are WP-specific and immediately remediable. Open-source (MIT), community-maintained, and extensible via custom rule packs.