|
SentinelMCP — Zero-Trust Security for AI Agents
The only MCP security gateway that detects tool poisoning, rug pulls, and shadow MCP servers in real time. Built for enterprise AI teams.
|
86 % |
|
RingZero — MCP & dependency security
Scan MCP configurations and software dependencies. Trace risks across large Gradle builds to the modules and teams that introduce them.
|
85 % |
|
MCPRadar — MCP Debugging & Security Scanner
The only tool that combines MCP server debugging and security scanning. Inspect tool calls, replay requests, and automatically detect prompt injection, path...
|
84 % |
|
MCP Scanner — Security Scanner for MCP Servers
Free open-source security scanner for Model Context Protocol (MCP) servers. Detects tool poisoning, prompt injection, rug pulls, and cross-origin attacks. Vi...
|
84 % |
|
MCP Verdict: the neutral registry for MCP servers
A historical registry of eight passing MCP filesystem evaluations, scored with a public behavioral-testing method.
|
84 % |
|
MCP Skills — Pre-Install Trust Layer for MCP Servers
The pre-install trust layer for MCP servers and AI skills. MCP Skills scores install risk, public trust pages, monitoring, badges, and API workflows across 1...
|
84 % |
|
MCPsight · Inspect an MCP server before you trust it
MCPsight is an open-source scanner for MCP servers. It catches prompt injection, watches local servers in a sandbox, and fails your build when a server chang...
|
84 % |
|
AgentJury — Tested by agents. Trusted by builders.
AgentJury tests MCP servers and agent skills with automated test agents. Reliability scores, compatibility matrices, and security analysis backed by evidence.
|
83 % |
|
Nox — The security scanner that understands your AI app
Open-source security scanner with first-class AI app security: prompt injection at the call site, embedding leakage, agent over-privilege, MCP hardening, cro...
|
83 % |
|
MCPhound
Local-first MCP security intelligence platform
|
83 % |
|
MCP Server Security Registry: 184 Detection Rules
Search MCP servers and evaluate their security posture. 184 detection rules with AST taint analysis, capability graph, and entropy-based secret detection. Fr...
|
83 % |
|
CheckMCP — audit & secure your MCP servers
Audit, monitor and gate any MCP server: an explainable MCP Score /100, drift alerts, and a gateway that blocks tool-poisoning before it reaches your agent.
|
83 % |
|
Sentinel MCP — Local MCP Security & Detection
Discover, fingerprint and defend every MCP server your AI agents can reach. Sentinel is the local, open-source EDR for MCP — zero cloud, read-only by default.
|
83 % |
|
ServerPal — server and uptime monitoring
An agent on every server reporting CPU, memory, disk, load and OS health, plus HTTP, TCP, ping, DNS and heartbeat checks. Alerts, incidents and branded statu...
|
82 % |
|
ray.run: one endpoint for every MCP tool
Build MCP servers with your coding agent, run them beside remote MCP, OpenAPI, npm, and container services behind one endpoint, and publish them to your cust...
|
82 % |
|
CraftedTrust - MCP server trust registry
Search public MCP servers, run a lightweight scan, and review practical trust signals before connecting an agent.
|
82 % |
|
ShiftLeft — MCP Gateway for Secure AI Tool Access
MCP Gateway enforces policy on every AI tool call. SBOM scanning, secrets detection, tool audit, and offline-capable signed policies — local stdio and HTTP....
|
82 % |
|
stillrunning — Supply chain security for teams without security teams
Supply chain security that stops North Korean malware before it runs. Static analysis, guard daemon, npm/pip intercept. Running 24/7 on real infrastructure.
|
82 % |
|
ToolTrust Scanner for MCP Servers and AI Agents
ToolTrust Scanner for MCP servers and AI agents. Scan tools for prompt injection, excessive permissions, and supply-chain risk, then review grades and findin...
|
82 % |
|
MCP Scanner — Scan your MCP servers before attackers do. · by codelake
Free security scanner for Model Context Protocol (MCP) servers. Finds auth bypass, injection, SSRF, hardcoded secrets and 30+ vulnerabilities in seconds — we...
|
82 % |
Detected social media networks