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
Zahaawii icon
Sec6

An MCP server providing AI assistant capabilities, dynamic tool execution, and data management for ChromaDB and an external blog platform.

Setup Requirements

  • ⚠️Requires `CHROMADB_API_KEY` environment variable (for ChromaDB cloud access)
  • ⚠️Requires `GOOGLE_API_KEY` environment variable (for embedding model, as per `compose.yaml`)
  • ⚠️Requires Docker and Docker Compose to run with the provided `compose.yaml`
Verified SafeView Analysis
The server correctly uses environment variables for API keys, avoiding hardcoded secrets. There's no use of 'eval' or other directly exploitable patterns. However, the 'Next Steps' section identifies current weaknesses in backend robustness, specifically 'Input validation' and 'Strong typed error responses'. For an AI-driven tool execution platform, a lack of explicit input validation in service methods for parameters passed to external APIs or the vector store (e.g., blog post content, ChromaDB document fields) poses a potential risk if the AI generates malformed or malicious input. Generic exception handling also limits clear error reporting.
Updated: 2026-01-10GitHub→
0
0
Medium Cost
mathisto icon
Sec9

Provides context-aware web fetching and content extraction tools for Large Language Models to efficiently process web content without exceeding token limits.

Setup Requirements

  • ⚠️Requires Python 3.10 or newer.
  • ⚠️Designed to be run as an MCP (Multi-tool Coordination Protocol) server, typically via `uvx`.
Verified SafeView Analysis
The server uses standard libraries (httpx, BeautifulSoup, lxml, tiktoken) for network requests, HTML parsing, and token counting. It implements basic rate limiting and request timeouts to prevent abuse and ensures timeouts are validated. Caching is in-memory only, reducing file-system related risks. No direct 'eval' or 'exec' calls, nor hardcoded sensitive information were found in the provided source code. The user agent is clearly identified.
Updated: 2025-11-28GitHub→
0
0
Medium Cost
5starsunited icon

neonpanel-mcp-server

by 5starsunited

Sec9

Facilitates secure, authenticated access to NeonPanel's business intelligence and operational APIs (inventory, finance, supply chain, import management) for AI models and external integrations via the Model Context Protocol.

Setup Requirements

  • ⚠️Requires a NeonPanel account and an OAuth2 client configured with NeonPanel's authorization server. Dynamic Client Registration (DCR) is supported via a helper script.
  • ⚠️Many tools require AWS Athena configuration with appropriate IAM roles (for querying and writing to Iceberg tables), and S3 buckets for query output, implying AWS account setup and permissions for Athena and S3.
  • ⚠️Requires Node.js and npm (or similar package manager) to run locally.
Verified SafeView Analysis
The server implements robust security practices, including Zod for input validation, explicit SQL string escaping (e.g., `sqlStringLiteral`, `sqlVarcharArrayExpr`) to prevent injection in Athena queries, `jsonwebtoken` for JWT validation using JWKS, `requireBearer` middleware for authentication, and `rateLimit` middleware. The `renderSqlTemplate` function is generic, but its callers (Athena tool registration functions) correctly sanitize inputs. There are no indications of 'eval' or malicious obfuscation. Hardcoded secrets are present only in test scripts, not in the main application logic.
Updated: 2026-01-18GitHub→
0
0
Medium Cost
JayaSamuthraDevi icon

mcp-server

by JayaSamuthraDevi

Sec9

A Model Context Protocol (MCP) server for secure access to billing and compute information, authenticated via Keycloak for AI assistant integration.

Setup Requirements

  • ⚠️Requires a running and configured Keycloak server instance with a specific client setup.
  • ⚠️Relies on the `uv` package manager for dependency management and running the server.
  • ⚠️Requires Python 3.14+ as per `pyproject.toml`.
Verified SafeView Analysis
The server uses OAuth2 via Keycloak for authentication, with sensitive configuration loaded from environment variables (no hardcoded secrets). It leverages httpx for secure asynchronous HTTP requests to upstream services. Upstream API keys and secrets are expected as claims within the verified OAuth token, shifting the security burden for these credentials to the Keycloak setup and token issuance process. No 'eval' or obvious malicious patterns found in the provided code. Requires HTTPS in production and careful handling of client secrets as emphasized in the README.
Updated: 2025-12-28GitHub→
0
0
Low Cost
sslotsky icon

chess-advice-mcp

by sslotsky

Sec8

A microservice component that provides strategic chess advice based on board state and move history using the OpenAI API.

Setup Requirements

  • ⚠️Requires OpenAI API Key (Paid)
Verified SafeView Analysis
Uses process.env for API key. Default cors() allows all origins, which might be too permissive depending on deployment context. Potential for prompt injection through user inputs (moveHistory, difficulty, tone) into the OpenAI prompt, though the prompt itself tries to constrain the LLM. No 'eval' or obfuscation found.
Updated: 2026-01-19GitHub→
0
0
Medium Cost

outline-mcp-server

by msasikumar

Sec8

Integrates a self-hosted Outline notes application with an MCP client for programmatic access and natural language interaction.

Setup Requirements

  • ⚠️Docker and Docker Compose required.
  • ⚠️Outline API Key required (from your self-hosted instance).
  • ⚠️Manual MCP client configuration involving Docker pathing and image tagging.
Verified SafeView Analysis
The primary security concern is the exposure of the Outline API key, which grants full access to the Outline instance. The README provides strong warnings and best practices for protecting this key, including not committing it to version control and considering Docker secrets. No 'eval' or obfuscation is apparent from the provided documentation.
Updated: 2025-11-17GitHub→
0
0
Low Cost
omarkamali icon

semango

by omarkamali

Sec2

A hybrid search engine for codebases, documentation, and knowledge bases, combining lexical (BM25) and semantic (vector) search with a web UI and REST API.

Setup Requirements

  • ⚠️Requires Go 1.23+, Node.js 20+, and Yarn for building from source.
  • ⚠️Requires CGO dependencies (FAISS, OpenBLAS, ONNX Runtime) to be manually installed or handled by provided scripts/Docker.
  • ⚠️Requires `OPENAI_API_KEY` environment variable if using OpenAI for embeddings (paid service), otherwise downloads local models (requires disk space and bandwidth).
  • ⚠️Configuration (semango.yml) must be initialized (`semango init`) and content indexed (`semango index`) before running the server.
Review RequiredView Analysis
CRITICAL VULNERABILITIES IDENTIFIED: 1. Authentication Bypass: The REST API (`/api/v1/search`, `/api/v1/health`, `/api/v1/stats`) is advertised as 'Token-authenticated' in the README, but the provided `api/server.go` code *does not implement any authentication middleware*. Requests from the `ui/App.tsx` also do not include an `Authorization` header. This means the API is publicly accessible without any token, which is a critical security vulnerability allowing unauthorized access to search and statistics. This server is unsafe to run in any exposed environment. 2. Arbitrary Code Execution (Plugins): The system supports dynamic loading of plugins (`.so` shared object files) specified in `semango.yml`. This feature allows arbitrary native code execution and presents a severe security risk if plugins are sourced from untrusted origins or if the configuration can be tampered with by an attacker. 3. External Binary Downloads: Installation scripts (`install_faiss.sh`, `install_onnxruntime.sh`) use `curl -L` to download external libraries (FAISS, ONNX Runtime) from GitHub releases. While GitHub is generally trusted, this introduces a dependency on external integrity and can be a supply chain risk. 4. CGO Usage: The project extensively uses CGO for FAISS and ONNX Runtime. This exposes the system to potential C/C++ vulnerabilities (e.g., buffer overflows, memory corruption) that Go's memory safety features typically mitigate. Careful review of CGO-bound code is essential.
Updated: 2025-12-14GitHub→
0
0
Medium Cost
vietld-3584 icon

autoTesting

by vietld-3584

Sec3

This server is designed for automated testing, likely related to the Minecraft Protocol (MCP.

Review RequiredView Analysis
No source code provided beyond the README, making a comprehensive security audit impossible. Cannot verify the absence of 'eval', obfuscation, hardcoded secrets, network risks, or malicious patterns. Therefore, a low safety score is assigned as a precaution.
Updated: 2026-01-19GitHub→
0
0
Low Cost
bioanywhere icon
Sec9

Provides Docker deployment configuration for an MCP (Managed Cloud Platform) server, including local development and CI/CD setup.

Setup Requirements

  • ⚠️Docker required
  • ⚠️Docker Compose recommended for orchestration
Verified SafeView Analysis
The provided code consists solely of Docker deployment configurations and scripts. It does not contain application logic. No 'eval' or obvious malicious patterns were found within these configuration files. The server exposes port 3000, which is standard. Data volume mounts './data' locally, which requires attention regarding sensitive data persistence if applicable. 'GITHUB_TOKEN' is mentioned for GitHub Container Registry authentication but is not hardcoded within the deployment configurations.
Updated: 2025-12-01GitHub→
0
0
Medium Cost
DMGoose icon
Sec9

Connects Notion workspaces to Claude Desktop, enabling Claude to search, read, and query Notion content with advanced features.

Setup Requirements

  • ⚠️Requires Node.js 24.11.0 (recommended via Volta)
  • ⚠️Requires a Notion account with admin access to create integrations and share content.
  • ⚠️Requires manual configuration within the Claude Desktop application to register the MCP server and set the NOTION_TOKEN environment variable.
  • ⚠️Notion integration needs explicit access to pages/databases you want Claude to access.
Verified SafeView Analysis
Notion API token is handled as an environment variable (NOTION_TOKEN), preventing hardcoding in the codebase. Input validation for tool parameters is performed using Zod schemas. No 'eval' or direct command execution with unsanitized user input is detected in the provided source code. Standard SDKs are used for Notion API integration and MCP communication, reducing direct exposure to common vulnerabilities. General dependency vulnerabilities remain a potential risk, as with any software.
Updated: 2025-11-19GitHub→
0
0
High Cost
cvidmar icon
Sec7

Provides Claude Code with an MCP server to get second opinions and code reviews from OpenAI's Codex CLI, acting as a bridge between the two AI assistants.

Setup Requirements

  • ⚠️Requires OpenAI Codex CLI to be installed and available in the system's PATH (validated on server startup).
  • ⚠️Requires Go 1.21+ to build the server executable.
  • ⚠️Requires manual configuration in Claude Code's JSON configuration file (`~/.config/claude/claude_code_config.json`) with the absolute path to the built executable.
Verified SafeView Analysis
The server uses `os/exec` to call external commands ('codex' and 'git'). While user-supplied arguments for `codex exec` are passed after a `--` delimiter, which helps prevent direct shell command injection from the prompt content, the security ultimately depends on the robustness of the underlying `codex` CLI in handling potentially malicious input within its prompt arguments. The `git diff` commands are executed with fixed arguments and do not directly incorporate user-controlled strings as command arguments, posing a lower risk. No hardcoded secrets or network listening are present in the server's own Go code.
Updated: 2025-12-03GitHub→
0
0
Medium Cost
davidpadbury icon

music-assistant-mcp

by davidpadbury

Sec9

Controls Music Assistant for multi-room audio, playback management, and music discovery.

Setup Requirements

  • ⚠️Requires a running Music Assistant server instance.
  • ⚠️Requires a long-lived access token from your Music Assistant server.
  • ⚠️Requires 'uv' (package manager) to be installed on the system.
  • ⚠️Requires Python 3.11 or newer.
Verified SafeView Analysis
The server uses Pydantic for input validation on all tool parameters, which helps prevent injection attacks. It relies on the `music-assistant-client` library for interaction with the Music Assistant server. No direct `eval` or `exec` calls are present in the provided source code, nor any direct system command execution without proper sanitization. Secrets (URL, token) are expected via environment variables, not hardcoded. The recommended installation method `uvx` runs the code in an isolated environment.
Updated: 2026-01-19GitHub→
PreviousPage 685 of 760Next