Stop Searching. Start Trusting.

The curated directory of MCP servers, vetted for security, efficiency, and quality.

Browse DirectoryWhy Us?

Tired of the MCP "Marketplace" Chaos?

We built MCPScout.ai to solve the ecosystems biggest pain points.

No Insecure Dumps

We manually analyze every server for basic security flaws.

Easy Setup

Our gotcha notes warn you about complex setups.

Avoid "Token Hogs"

We estimate token costs for cost-effective agents.

Products, Not Demos

We filter out "Hello World" demos.

CATEGORIES:
SORT:

Vetted Servers(9120)

0
0
Low Cost
Loctree icon

rmcp_memex

by Loctree

Sec8

A local Model Context Protocol (MCP) server providing Retrieval-Augmented Generation (RAG) capabilities with embedded vector storage and local embeddings/reranking.

Setup Requirements

  • ⚠️Requires Rust toolchain with Cargo (stable)
  • ⚠️Requires Protobuf compiler at build time
  • ⚠️Optional MLX HTTP server required for MLX-based embeddings/reranking (otherwise falls back to fastembed)
Verified SafeView Analysis
The server runs as a local process, communicating via stdin/stdout using JSON-RPC, which inherently limits direct network attack surface. It relies on well-established Rust libraries for HTTP (reqwest), data serialization (serde_json), and vector storage (lancedb). No 'eval' or similar dynamic code execution patterns were found. Environment variables are used for configuration, preventing hardcoded secrets. Processing external files (PDFs via pdf-extract, arbitrary text) does introduce a potential risk if these libraries or the parsing logic are vulnerable to malformed inputs. The MLX bridge relies on a separate, assumed-to-be-trusted, local MLX HTTP server.
Updated: 2026-01-11GitHub→
0
0
Medium Cost
hsiangjenli icon

mcp-tw-gpss

by hsiangjenli

Sec8

An AI-ready Model Context Protocol (MCP) server for searching patents in the Global Patent Search System (GPSS).

Setup Requirements

  • ⚠️Requires GPSS API access code set as `USER_CODE` environment variable (access to GPSS API is likely paid or restricted).
  • ⚠️Requires Python 3.12+ for local execution.
  • ⚠️Docker is recommended for easier deployment and environment consistency.
Verified SafeView Analysis
The server correctly retrieves the `USER_CODE` for API authentication from environment variables, preventing hardcoded secrets. It makes external HTTP requests to the GPSS API, which is inherent to its function. While Pydantic schemas provide input validation for the tool's parameters, the handling of `keywords` for potential injection into the external GPSS API relies on the GPSS API's own sanitization, which is a common pattern for API wrappers but an external dependency for security.
Updated: 2025-11-26GitHub→
0
0
Medium Cost

DevBuddy

by dazinator

Sec9

Analyzes .NET codebases and executes shell commands in a secure, containerized environment via the Model Context Protocol (MCP) for AI agents.

Setup Requirements

  • ⚠️Docker Desktop required for server deployment
  • ⚠️.NET 8.0 SDK required for local development or generating local dev certificates
  • ⚠️Claude Desktop requires specific configuration (Docker bridge or remote connector)
  • ⚠️Git Personal Access Tokens (PATs) are needed for remote Git operations
Verified SafeView Analysis
The server features robust security, including direct process spawning to prevent command injection, comprehensive path validation, configurable command allow/deny lists, and enforced timeouts. Container hardening utilizes non-root user execution, dropped capabilities, and no-new-privileges. Error messages are sanitized for production, and detailed audit logging with sensitive data redaction is implemented. The default `docker-compose.yml` uses a read-write workspace for development, which is a known flexibility for development but is clearly noted to be read-only (`:ro`) for production deployments.
Updated: 2025-11-27GitHub→
0
0
High Cost

taobao_mcp

by JeremyDong22

Sec9

Scrapes comprehensive product information from Taobao/Tmall for AI assistants to perform product research, comparison, and analysis.

Setup Requirements

  • ⚠️Requires manual configuration of absolute paths in `.mcp.json` for both the Python interpreter (preferably venv) and `server.py`, which is a common source of setup errors.
  • ⚠️Mandatory initial QR code login to Taobao/Tmall through a manually launched browser window, requiring user interaction to establish a persistent session.
  • ⚠️Requires Playwright's Chromium browser to be separately installed via `playwright install chromium` after Python dependencies are met.
Verified SafeView Analysis
The server uses Playwright to launch a Chromium browser for web scraping, including handling user logins via QR code, which means it manages session cookies and interacts with external websites. It is designed to run locally as an MCP server, reducing external attack surface. Transparency in login process (QR code) and local execution contributes to a high safety score, assuming the user trusts the repository code.
Updated: 2025-11-17GitHub→
0
0
Medium Cost
kipkemoifred icon

mcp_server

by kipkemoifred

Sec9

A Spring Boot application demonstrating Spring AI tool integration for retrieving weather information and a GraphQL client for fetching country data, acting as a service component for conversational AI.

Setup Requirements

  • ⚠️Requires Java Development Kit (JDK) 17+.
  • ⚠️Building the application requires a build tool like Maven or Gradle (not explicitly provided in source, but implied by project structure).
Verified SafeView Analysis
The application makes external API calls to 'https://countries.trevorblades.com/' (GraphQL) and 'https://api.weather.gov/' (REST). No hardcoded sensitive secrets are visible. There are no signs of 'eval' or other dangerous dynamic code execution patterns. The 'User-Agent' in WeatherService includes a placeholder email ('your@email.com') which should be configured for a production environment but isn't a direct security risk.
Updated: 2025-11-25GitHub→
0
0
Low Cost
dmj03 icon
Sec9

Automate Jira issue creation and management through an MCP server.

Setup Requirements

  • ⚠️Requires access to a Jira instance and a user with permissions to create issues and generate an API token.
  • ⚠️The environment variables JIRA_URL (optional, defaults to deepakjohnart.atlassian.net), JIRA_EMAIL, and JIRA_API_TOKEN must be correctly configured.
  • ⚠️Requires Python packages 'fastmcp' and 'httpx' to be installed.
Verified SafeView Analysis
Uses environment variables for sensitive credentials (JIRA_EMAIL, JIRA_API_TOKEN), which is a good security practice. No 'eval' or other dynamic code execution detected. The server connects to an external Jira API, which implies a reliance on the security of the configured Jira instance and API token.
Updated: 2026-01-19GitHub→
0
0
Low Cost
NirjharDeb icon
Sec3

An MCP server prototype for evaluating hidden-payload delivery to LLMs in agentic IDEs for ML security research and defensive testing.

Setup Requirements

  • ⚠️Requires `fastmcp` Python library (implicit dependency)
  • ⚠️Requires Python 3.8+
Verified SafeView Analysis
The project's explicit goal is to evaluate 'hidden-payload delivery to LLMs', and its source code implements several techniques that are inherently security-relevant patterns. The `translate_text` tool, for instance, actively injects a base64-encoded 'SYSTEM INSTRUCTION' into its output, which contains instructions intended to manipulate the LLM's behavior (e.g., to call another tool). Other tools use zero-width characters, hidden HTML comments/divs, and ANSI escape codes for obfuscation. The `read_gdoc` tool performs network requests to external Google Docs, which, while targeted, represents a potential vector for SSRF if not carefully managed or if a malicious URL could be injected. While the project states 'no malicious use' and is for 'defensive testing', the techniques demonstrated are direct implementations of methods used in LLM instruction injection and payload delivery, which are security vulnerabilities in a general context. The `send_email` tool is simulated and does not actually send emails, which is a positive from a security perspective.
Updated: 2025-11-30GitHub→
0
0
Medium Cost
PaulRoze icon
Sec9

Provides comprehensive documentation, code patterns, and troubleshooting guides for NVIDIA ConnectX-7, DOCA, VMA, and RDMA driver development, primarily for high-performance networking and broadcast/media teams.

Setup Requirements

  • ⚠️Requires Python >=3.10.
  • ⚠️Manual setup involves cloning a Git repository, creating a virtual environment, installing dependencies, and configuring a specific absolute path to the executable in Claude Code's settings.json.
  • ⚠️Requires internet access to fetch documentation if not already cached.
Verified SafeView Analysis
The server fetches external web content using `httpx` and parses it with `BeautifulSoup` and `markdownify`. While the documentation sources are hardcoded, reducing arbitrary URL fetching risk, parsing external content can always have edge cases. No use of `eval` or obvious malicious patterns found. File system access is limited to a cache directory for downloaded documentation.
Updated: 2026-01-19GitHub→
0
0
Low Cost
bioanywhere icon

ui

by bioanywhere

Sec8

API client for accessing and processing world news from various sources.

Setup Requirements

  • ⚠️Requires API keys (api-key and/or x-api-key) which must be obtained from worldnewsapi.com.
  • ⚠️The C++ client requires CMake 3.2+, Qt, and a C++ compiler.
  • ⚠️The Python client requires Python 3.7+.
Verified SafeView Analysis
The project consists of OpenAPI-generated client libraries across multiple languages. The C++ client includes an OAuth implementation that starts a local TCP server on port 9999 to handle redirect URIs. While this is a standard pattern for some OAuth flows, exposing a local port could be a potential, albeit low, attack surface if not properly secured (e.g., binding only to localhost and correctly validating incoming requests). API keys are expected to be provided by the user programmatically; the library itself does not hardcode them. No obvious 'eval' or obfuscation techniques were detected.
Updated: 2025-12-13GitHub→
0
0
Medium Cost
asaprivate icon

gitflow-for-pms

by asaprivate

Sec9

Enables non-technical Product Managers to manage Git and GitHub workflows directly through AI coding assistants.

Setup Requirements

  • ⚠️Requires Docker Desktop (or Rancher Desktop) to run the PostgreSQL database and Redis cache.
  • ⚠️Requires Node.js v20+ to run the server.
  • ⚠️Requires manual creation and configuration of a GitHub OAuth App to obtain GITHUB_CLIENT_ID and GITHUB_CLIENT_SECRET, which is a multi-step process outside the direct installation commands.
Verified SafeView Analysis
The server demonstrates strong security practices: GitHub OAuth tokens are stored in the system keychain via 'keytar' (or encrypted in DB as fallback), and are explicitly never logged in plaintext. Database queries use parameterized statements to prevent SQL injection. Input validation is performed using 'zod' for all MCP tool arguments. The 'GitService' correctly sanitizes URLs before logging to prevent token exposure. The 'PolicyRejectionHandler' actively detects and helps users mitigate GitHub Push Protection violations (e.g., exposed secrets) by guiding safe history sanitization and re-pushing using '--force-with-lease'.
Updated: 2026-01-17GitHub→
0
0
High Cost
igorrodrigz icon

MCP-SERVER

by igorrodrigz

Sec6

A FastAPI backend serving as a virtual tourist guide chatbot, leveraging the Groq API for natural language processing and integrating with a database to provide location-specific information for Uruguaiana, Brazil.

Setup Requirements

  • ⚠️Requires a Groq API Key, implying a dependency on a potentially paid external AI service.
  • ⚠️Requires a database setup: by default, it expects a MySQL instance with pre-existing tables (`pontos_turisticos_pontoturistico`, `estabelecimentos_restaurante`). If configured for SQLite, a `turismo.db` file with `cidades`, `pontos_turisticos`, and `restaurantes` tables is needed. The database schema itself is not provided in the source code.
  • ⚠️External Python dependencies (e.g., fastapi, uvicorn, requests, python-dotenv, mysql-connector-python) must be installed prior to running.
Verified SafeView Analysis
The application correctly utilizes environment variables for sensitive API keys and database credentials, which is a good security practice. However, the `CORSMiddleware` is configured with `allow_origins=["*"]`, which permits requests from any domain. This setting is generally unsafe for production environments as it exposes the API to cross-origin attacks and should be restricted to specific, trusted origins.
Updated: 2025-12-10GitHub→
0
0
Medium Cost
premkalyan icon

jira-mcp

by premkalyan

Sec9

Integrates Jira Cloud with AI agents and automation systems via Model Context Protocol, providing comprehensive issue, project, and sprint management capabilities.

Setup Requirements

  • ⚠️Requires Node.js 18.0.0 or higher.
  • ⚠️Requires access to a Jira Cloud instance with an API Token and appropriate permissions (Browse projects, Create/Edit issues, Work on issues, Add comments).
  • ⚠️For multi-tenant Vercel deployment, it depends on an external Project Registry (https://project-registry-henna.vercel.app) and requires setting `REGISTRY_URL` and `REGISTRY_AUTH_TOKEN` environment variables in Vercel.
Verified SafeView Analysis
The project has explicitly addressed and fixed a critical security vulnerability involving a hardcoded Bearer token by moving it to an environment variable (`REGISTRY_AUTH_TOKEN`). Jira API credentials are dynamically fetched from a Project Registry per request in the serverless deployment, preventing their direct exposure. Input validation is implemented for environment variables and Jira configuration fields. Rate limiting is present to prevent API abuse. While documentation might suggest a default `REGISTRY_AUTH_TOKEN` value, the code itself reads from environment variables, which is a secure practice. Rotating this registry token is a deployment best practice.
Updated: 2025-12-14GitHub→
PreviousPage 699 of 760Next