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
wminson icon
Sec8

Enables AI assistants to interact with GitHub Copilot CLI for command suggestions, explanations, and configuration management through a standardized Model Context Protocol (MCP) interface.

Setup Requirements

  • ⚠️Requires GitHub Copilot subscription (paid)
  • ⚠️Requires GitHub CLI (gh) installed
  • ⚠️Requires GitHub Copilot extension for gh (`gh extension install github/gh-copilot`)
  • ⚠️Requires Python 3.10+
Verified SafeView Analysis
The server safely executes GitHub CLI commands via subprocess.run, passing arguments as a list to mitigate shell injection risks. No eval or hardcoded secrets are present. The SSE transport (server_sse.py) defaults to binding on 0.0.0.0 and attempts to disable certain transport security features (DNS rebinding protection, host/origin restrictions) for broad network accessibility. While this enhances flexibility, it requires careful deployment considerations regarding network exposure.
Updated: 2025-12-11GitHub→
0
0
Medium Cost
theflysurfer icon

hydraspecter

by theflysurfer

Sec7

Multi-headed browser automation for web scraping, testing, or data collection with robust anti-detection and session persistence capabilities.

Setup Requirements

  • ⚠️Requires Node.js 18.0.0 or higher.
  • ⚠️Windows users requiring Chrome session syncing (VSS) need to run a PowerShell script (`sync-cookies-vss.ps1`) with Administrator privileges once.
  • ⚠️Chrome v127+ App-Bound encryption may prevent automatic cookie transfer from your real Chrome profile to HydraSpecter's `pool-0`. Manual login into HydraSpecter's `pool-0` (visible mode) and subsequent syncing to other pools via `scripts/sync-pools.ps1` will be necessary for persistent sessions on such domains.
  • ⚠️The default headless mode for anti-detection features is `false` (visible browser), which can consume more system resources.
Verified SafeView Analysis
The server exposes powerful primitives like `browser_evaluate` for arbitrary JavaScript execution within the browser context, which is standard for browser automation but requires careful handling by the orchestrating model to prevent misuse. The `trySyncViaVSS` function on Windows uses `execSync` to run a PowerShell script with string interpolation for directory paths. While these paths are derived internally from `os.homedir()`, it could pose a risk if path variables were somehow compromised to allow command injection. The tool also provides network access for navigation and proxying, which is inherent to its functionality.
Updated: 2026-01-19GitHub→
0
0
High Cost
SPerekrestova icon

GitHub_MCP_Server

by SPerekrestova

Sec9

Provides Model Context Protocol (MCP) access to GitHub documentation via API for AI agents.

Setup Requirements

  • ⚠️Requires a GitHub Personal Access Token with 'repo', 'read:org', and 'read:user' scopes.
  • ⚠️Docker must be installed, or a Python environment with dependencies must be set up.
  • ⚠️Requires manual configuration in Claude Desktop's 'claude_desktop_config.json' file.
Verified SafeView Analysis
The code itself is clean, handles secrets via environment variables, and makes standard API calls to GitHub. No 'eval' or similar dangerous functions are used. The primary security consideration is the broad scope required for the GitHub Personal Access Token ('repo', 'read:org', 'read:user'), which grants significant access to private repositories if misused outside the application's read-only tools.
Updated: 2026-01-16GitHub→
0
0
Medium Cost
rexnzm icon
Sec3

Retrieval-Augmented Generation (RAG) server for multi-format document ingestion and vector database integration, supporting Ollama or OpenAI embeddings.

Setup Requirements

  • ⚠️Requires Python >= 3.13 (as per pyproject.toml).
  • ⚠️Requires either an OpenAI API Key (which is a paid service and needs internet access, set via OPENAI_API_KEY) OR a local Ollama server running (which requires installing Ollama, downloading an embedding model like nomic-embed-text, and approximately 4GB of RAM, set via OLLAMA_BASE_URL and OLLAMA_EMBED_MODEL).
  • ⚠️The official setup guide in the README (downloading executables from raw GitHub links) is a critical security risk and should not be followed.
Review RequiredView Analysis
The primary security concern arises from the distribution method outlined in the README, which instructs users to download and run executables (.exe, .dmg) or archives (tar.gz) directly from raw GitHub URLs. This bypasses secure software distribution practices and poses a significant supply chain risk, making it unsafe to run by following the project's own setup instructions. Additionally, the 'ingest_document' tool can download files from arbitrary URLs, which could be exploited for denial-of-service or malicious content download if the server were exposed to untrusted inputs. However, no 'eval' or explicit hardcoded secrets are found in the provided Python source code, and network calls are to standard endpoints for Ollama or OpenAI.
Updated: 2026-01-19GitHub→
0
0
Low Cost
Sec7

A server-side application designed for taking and managing notes, likely providing an API for client applications.

Setup Requirements

  • ⚠️Requires Java Development Kit (JDK) installed to build and run.
  • ⚠️Likely requires a database (e.g., H2, MySQL, PostgreSQL) setup and configuration for data persistence.
  • ⚠️No README file; detailed setup instructions, configuration specifics, and API documentation will need to be inferred from the `pom.xml` and source code.
Verified SafeView Analysis
No source code provided for detailed analysis. Standard Java/Maven project structure suggests general robustness. As a server application, it inherently involves network interfaces, requiring proper security practices (e.g., input validation, authentication, authorization) which cannot be assessed without code review. No obvious signs of 'eval' or obfuscation from the file list.
Updated: 2025-11-17GitHub→
0
0
Low Cost
AnuragRanjan2003 icon

NotesMCPServer

by AnuragRanjan2003

Sec7

This server provides a simple API for managing notes, allowing users to add, retrieve, update, and delete notes in an in-memory storage.

Setup Requirements

  • ⚠️Requires Python 3.13+
  • ⚠️Requires 'mcp[cli]' package (version >= 1.22.0)
Verified SafeView Analysis
The server uses in-memory storage, which limits direct database injection risks. However, there's no explicit input sanitization shown for string inputs (title, body) at the application layer. While the 'FastMCP' framework might offer some protection, reliance on it without explicit checks could be a risk if a persistent database were integrated later. No 'eval' or hardcoded secrets found. The concurrent utilities (ConcurrentDict, ConcurrentCounter) are basic thread-safe wrappers, not security concerns themselves.
Updated: 2025-11-22GitHub→
0
0
Low Cost
saifeezibrahim icon

aws-mcp-server-devops-ai

by saifeezibrahim

Sec9

Enables AI assistants to execute AWS CLI commands and access AWS resource context via the Model Context Protocol (MCP).

Setup Requirements

  • ⚠️Requires AWS credentials configured on the host machine (via ~/.aws files or environment variables).
  • ⚠️Docker is the strongly recommended deployment method for security and reliability (or Python 3.13+ and AWS CLI installed locally).
  • ⚠️Requires `uv` or `pip` for dependency management if running natively.
Verified SafeView Analysis
The server employs a robust multi-layered command validation system (allow/deny lists, regex rules, Unix command allowlist in pipes) to prevent dangerous AWS CLI operations and command injection. It uses `asyncio.create_subprocess_exec` with `shlex.split` for safe command execution. No hardcoded secrets. Relies on host-provided AWS credentials. Strong emphasis on Docker isolation and IAM least privilege in documentation. The default `stdio` transport is secure; `sse` requires careful network exposure. The code explicitly prevents many common AWS security misconfigurations and dangerous commands.
Updated: 2025-11-26GitHub→
0
0
Low Cost
carloshenriquecarniatto icon

Code-Tools

by carloshenriquecarniatto

Sec9

A set of developer utilities for analyzing .NET codebases, including dead code detection and other analysis tools.

Setup Requirements

  • ⚠️Requires .NET 10.0 SDK for the main application.
  • ⚠️Requires Node.js 20+ and npm for CI/CD and release automation tooling.
  • ⚠️The release workflow requires a GitHub token with 'contents: write' permissions, which may conflict with strict branch protection rules (solutions are provided in README).
Verified SafeView Analysis
The core application is a local .NET console tool with no apparent network vulnerabilities or hardcoded secrets in the provided snippets. The CI/CD process involves JavaScript-based tooling (`semantic-release`, `commitlint`) and a Node.js script (`tools/update-csproj.js`) for updating the `.csproj` file. The `update-csproj.js` script uses regex-based string manipulation for file modification, which is noted as potentially fragile, though not a direct security exploit in this context. The main security consideration is the `GITHUB_TOKEN` or `PERSONAL_TOKEN` permissions required for the release workflow to write to the repository, which is a standard and well-documented concern for automated CI/CD releases.
Updated: 2025-11-25GitHub→
0
0
Low Cost
Mist-wu icon

blbl-mcp-server

by Mist-wu

Sec9

This MCP server provides a tool to search for users on Bilibili using the Bilibili API.

Setup Requirements

  • ⚠️Python 3.12+ required
Verified SafeView Analysis
The code primarily acts as a wrapper around the `bilibili-api-python` library, exposing a search function. There are no 'eval' or 'exec' calls, hardcoded secrets, or obvious malicious patterns. The input 'keyword' is passed directly to a library function, which is generally considered safe. The server runs via standard I/O, limiting direct network exposure beyond the Bilibili API calls themselves.
Updated: 2025-11-19GitHub→
0
0
Low Cost
johanbolofsson icon

mcp-server-test

by johanbolofsson

Sec9

This server exposes historical Champions League winners data via an MCP (Model Context Protocol) interface, allowing AI clients like Claude for Desktop to query sports results.

Setup Requirements

  • ⚠️Requires .NET 8 SDK or higher
  • ⚠️Requires an MCP client (e.g., Claude for Desktop) for interaction
  • ⚠️Manual configuration of the absolute project path is needed in the MCP client's config file
Verified SafeView Analysis
The server appears to be designed for local execution, serving static data from a local JSON file. There are no explicit indications of network risks beyond standard local server operations, 'eval' usage, hardcoded secrets, or malicious patterns in the provided source code and README. It relies on standard .NET framework practices. The primary risk would be potential vulnerabilities in the .NET runtime or third-party libraries, which is common for any application.
Updated: 2025-12-02GitHub→
0
0
Low Cost
yadavatul868 icon

test-remote-mcp-server

by yadavatul868

Sec9

This server provides a simple remote API for basic arithmetic operations and random number generation using the FastMCP framework.

Setup Requirements

  • ⚠️Requires Python 3.13 or newer
Verified SafeView Analysis
The server exposes simple arithmetic and random number generation functions. There is no use of 'eval', 'exec', or other inherently dangerous functions that could lead to arbitrary code execution. It listens on '0.0.0.0' by default, which means it is accessible from all network interfaces; ensure appropriate firewalling if exposed to untrusted networks. No hardcoded secrets or sensitive information are identified within the provided source code.
Updated: 2025-11-24GitHub→
0
0
Medium Cost
alexcpn icon
Sec6

Provides an MCP server for code review, offering tools to analyze Python, Go, and C++ GitHub repositories via AST parsing.

Setup Requirements

  • ⚠️Requires `uv` package manager.
  • ⚠️Requires local git client installed.
  • ⚠️Requires Python version 3.10 or 3.11.
  • ⚠️Requires internet access to clone GitHub repositories.
Review RequiredView Analysis
The server clones arbitrary GitHub repositories provided by the `github_repo` parameter. While cloning occurs in a temporary directory and only shallow clones (`depth=1`) are performed, there is an inherent risk that a malicious repository could contain `.git` hooks that execute during the clone operation. The server then parses the code but does not directly execute it.
Updated: 2025-12-19GitHub→
PreviousPage 652 of 760Next