MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
Model Context Protocol (MCP) server providing comprehensive cybersecurity intelligence from HUMAN Security. Offers real-time attack monitoring, threat detection, fraud prevention, PCI DSS compliance validation, and supply chain security for AI-powered app
npm MIT -
Security memory with evidence attached. Human review, isolated client memories, and auditable recall in your own Cloudflare account. Built by Agent9.dev
MIT ★ 5 -
إدارة وتنظيم آمن للملفات عبر خادم MCP مع تأكيد الحذف البشري | A secure local MCP server for file triage, organization, and human-in-the-loop safe deletions.
NOASSERTION -
Agent-native web security: a human and an AI agent inspect and fix an authorized site together, via six WebMCP tools on document.modelContext. Passive-only, human-approved. Built for the OpenAI WebMCP Challenge.
MIT -
agentshieldgithub.com/sohailshaikh20/agentshieldupdated this year
Identity & access security for non-human identities. Open-source, starting with AI agents and MCP.
npm MIT -
🦭 Secure-by-design AI-agent backend. Isolated by Postgres RLS, gated by human approval, pinned to its data residency. FastAPI · PydanticAI · MCP.
MIT ★ 1 -
Model Context Protocol (MCP) server providing comprehensive cybersecurity intelligence from HUMAN Security. Offers real-time attack monitoring, threat detection, fraud prevention, PCI DSS compliance validation, and supply chain security for AI-powered app
npm MIT -
webmcp-sentinelgithub.com/guyle/webmcp-sentinelupdated this year
Server-enforced authorization, human approval, and audit for AI-agent actions — a WebMCP security reference implementation
MIT -
Secure AI agent execution with MCP, deterministic policies, human approval, idempotency and auditable simulated financial operations.
MIT -
Security-first Binance MCP. 26 tools, human-confirmed spot trades, offline quant research.
pypi -
Binance Spot (security-first) + crypto analystgithub.com/Alejandrehl/kainext-binance-mcpno update data
Security-first Binance Spot MCP + honest crypto analyst. 25 tools, 5 prompts, human-gated orders.
pypi -
AISRF, the AI Security & Research Framework: every LLM request from your apps and agents becomes a ticket a human approves before it reaches the backend. Analyzers, guardrails (Rebuff, LLM Guard, NeMo, Lakera), red teaming (garak, promptfoo, PyRIT), source code review, reports, MCP server.
Apache-2.0 ★ 2 -
AgentGuard is an open-source, local-first MCP security gateway that protects AI agents with policy-based tool permissions, human approvals, audit logs, and policy simulation. Developed by Belfeki.tn.
Apache-2.0 -
A security-first TypeScript/Node.js MCP server for simulated AWS, Azure, and GCP infrastructure, demonstrating policy enforcement, human approval workflows, auditability, and safe agent-driven operations.
MIT -
sideguardgithub.com/alisaitteke/sideguardupdated this year
MCP guard & vibe coding security tool for Cursor and Claude Code. Fail-closed human-in-the-loop approval for shell commands and MCP tools—local daemon, YAML policy, no cloud.
MIT ★ 2 -
sentricore-agent-securitygithub.com/Pothan0/sentricore-agent-securityupdated this year
Security gateway for AI agents on REST and MCP: prompt-injection defense, tool governance, human-in-the-loop escalation, tamper-evident audit logs.
-
Collaborative application security testing between humans and agents via CLI and MCP
go MIT ★ 43 -
lazysshgithub.com/MauricioPerera/lazysshupdated this year
Secure dual-interface SSH: a human TUI plus agent-safe CLI and MCP tools.
-
Offensive-security tools, skills, and MCP servers for local AI agents — privacy-first, small-model-driven. Authorized pentest/red-team companion: 6 MCP servers (27 tools), all irreversible classes human-gated.
Apache-2.0 -
One-command website security audit for humans and AI agents — scored 0-100, every finding explained with its fix, known CVEs matched via NVD. CLI + MCP server, zero dependencies.
MIT ★ 1 -
arx-mcp-servergithub.com/gethammerpath/arx-mcp-serverupdated this year
ARX MCP Server - Execute 100+ security operations with policy enforcement, audit logging, and human approvals
pypi MIT License -
Claude Code plugin for protect-mcp — MCP security gateway with per-tool policies, Ed25519-signed receipts, and human approval gates
MIT -
Policy checkpoint between AI agents and PayPal refunds: MCP security proxy, LangGraph risk pipeline with human-in-the-loop approval, adversarial evals.
MIT ★ 1 -
Security-first MCP filesystem server enabling Claude Desktop and AI tools to safely read, write, and manage local files with human-in-the-loop approval workflows
cargo MIT -
MCPgithub.com/Diluksha-Upeka/MCPupdated this year
Production-ready Enterprise MCP Operations Console featuring secure AI agent workflows, human-in-the-loop approvals, hybrid retrieval, and enterprise-grade authentication.
MIT -
Safe-by-default MCP server for GitHub Enterprise — orgs, repos, teams, rulesets, security alerts & Actions, with a security-posture baseline and governance controls (read-only by default, org-scoped, audited, human-confirmed).
npm MIT -
An MCP server and secure input broker: an AI agent can place a credential without ever receiving it, while a trusted human-controlled UI authorizes where it goes.
npm Apache-2.0 ★ 2 -
Human approval requests for agent actions: expiring, single-use decisions. Trusted executors must enforce the gate; not a same-user security boundary.
npm MIT -
Fast, secure SSH & SFTP Model Context Protocol (MCP) server in Rust with interactive TUI, Human-in-the-Loop (HITL) guardrails, and Pi Coding Agent integration.
MIT ★ 1 -
LineBreak security gate at the git/CI boundary: dependency CVE scan + AI SAST, human-approved overrides, git-native audit records
pypi Apache-2.0 -
Secure file exchange for AI agents — scanned, provenance-stamped files behind expiring capability URLs, with human approval gates and Ed25519-signed receipts. MCP server + JS/Python SDKs.
npm MIT -
@agentlock/mcp-servergithub.com/edies76/agent-lockerupdated this year
Agent-Lock MCP Server: Governance & security layer for Claude Desktop. Intercepts every tool call with Gemini risk analysis and Telegram human-in-the-loop approval.
npm MIT -
Evidence-backed security alert triage on LangGraph — picks the next check by information gain, every verdict ships with a verifiable proof. MCP server + Agent Skill, OpenTelemetry, human-in-the-loop.
MIT ★ 1 -
Secure MCP and automation hub for CareerOps, integrating n8n and OpenClaw with durable workflows, human review, PostgreSQL and an AI Agent Engine
MIT -
vantagate-mcp-servergithub.com/Aderix/vantagate-mcp-serverupdated this year
Human-in-the-Loop authorization gateway for AI Agents. Securely pause MCP workflows and route high-risk actions to human approvers via Slack or Email.
npm MIT ★ 1 -
Advisory NixOS change-review + CVE/KEV attestation gate for AI agents (MCP server). Grades security-relevant config diffs, attests closures against the CISA KEV catalog, gates on a human, keeps an append-only audit ledger.
MIT -
Agentic multi-cloud FinOps, security auto-remediation and AI observability platform. FastAPI + LangGraph agent swarm, corrective RAG (Qdrant), PyTorch CNN anomaly detection, Groq + Gemini LLM gateway, MCP server, human-approved Terraform fixes, Docker..
MIT ★ 3 -
Reliable AI agents over legacy enterprise databases: secure, observable and measured. MCP server with an AST SQL guard, independent verifier, SWAR detector for silent errors, human-approved writes, pre-registered evals with a CI gate, and an AWS CDK deployment.
MIT -
pundi-wallet-mcpupdated this year
Pundi Agentic Wallet MCP server — give AI agents secure, human-approved blockchain wallet capabilities
npm MIT -
An interactive, full-stack AI filesystem chatbot powered by Google Gemini and Model Context Protocol (MCP). Built with Python FastMCP, FastAPI, and React, it lets you safely manage files via natural language with real-time file tree sync, strict sandbox security, and human-in-the-loop deletion confirmation.
-
Secure local environment management for humans, CI, and AI agents with CLI, TUI, and MCP interfaces.
MIT -
An MCP server with a real security model and the tests that prove it: three permission tiers with human approval for destructive tools, a path jail, tool results marked as untrusted data, rate limits, a hash-chained audit log and a 47-check protocol conformance suite. 974 tests, no network.
MIT -
Argus is an agentic smart-contract security copilot for auditors. Deterministic static analysis, SWC-mapped retrieval, false-positive filtering, remediation diffs, and human-in-the-loop triage. TypeScript engine, Next.js workbench, MCP tool host. Educational samples only; never connects to a chain. Vercel is demo-mode (no API key required).
MIT -
Hermes-native browser hands for AI agents: persistent sessions, human-like actions, and secure uploads over MCP
npm MIT -
Tenant-safe legal document assistant on AWS Bedrock AgentCore: grounded RAG with verifiable citations, secure Gateway/MCP tools, and asynchronous document extraction with evidence-based human review.
-
MCP_Servergithub.com/Priya-1514/MCP_Serverupdated this year
A secure and interactive human verification system. Users log in with credentials and verify their identity by clicking randomly generated colored dots, making it harder for AI bots to bypass. An innovative alternative to traditional CAPTCHAs.
-
Multi-tenant platform for running AI agents against customer systems: MCP tools with human approvals, Postgres row level security, inspectable retrieval, and Kubernetes delivery.
-
A self-hosted unified inbox for humans and AI agents, with IMAP/POP3, SMTP, MCP, and a secure send API.
★ 3 -
Security-first Binance MCP server and multi-agent intelligence platform. Read market data, quant analytics, and propose spot trades behind human confirmation.
MIT ★ 4 -
AI-BO-Agentgithub.com/zunairaiqbal078/AI-BO-Agentupdated this year
A production-ready Human-in-the-Loop AI Business Operations Assistant (Next.js, Prisma, PostgreSQL) utilizing Model Context Protocol (MCP) to analyze project health and securely propose tasks under strict RBAC constraints.
-
OrderBridge-Enterprise-AI-MCP-Projectgithub.com/bwartman2015/OrderBridge-Enterprise-AI-MCP-Projectactive
AI-enabled supply chain integration hub: X12 EDI, FastAPI, Postgres, Redpanda outbox, and a secure MCP server with RBAC, audit trail and human approval
-
Agentic backend service (FastAPI + LangGraph + MCP) with an orchestrator routing to specialized agents, a 3-node agentic loop, security-guarded MCP tools (SQL/SSRF/path-traversal checks), and human-in-the-loop approval for sensitive operations.
MIT -
agentic-sdlc-mcp is an SDLC orchestration layer and control plane built on the Model Context Protocol (MCP). It wraps GitHub APIs into high-level, opinionated tools that enforce traceability, human-in-the-loop gates, quality thresholds, and security checks for AI coding agents.
npm MIT ★ 1 -
Agent-neutral security checks before you ship — SAST, secrets, and dependency audit as a merge gate. Find, understand, and fix vulnerabilities in AI- or human-written code. BYO-inference, $0, MIT.
pypi MIT -
@agentopolis/sapientdb-mcpupdated this year
SapientDB is an agent-first, MCP-native markdown database where humans and AI agents collaborate on shared documents. Text is stored as individually addressable, permissioned logical text units, enabling secure block-level reads and writes without loading
npm MIT -
firesharkgithub.com/HendrikReh/firesharkupdated this year
Rust packet analyzer for LLMs and humans — 11-protocol dissection, 9-command CLI, 21-tool MCP server for AI-driven security audits, Wireshark-style display filters, stream tracking, and optional tshark backend
Apache-2.0 ★ 1 -
AI agent that browses any website in a real browser (Playwright + Gemini/Claude/OpenAI), with security enforced in code: SSRF protection, network guard, secret redaction, human approval, MCP server and end-to-end evals.
MIT -
An enterprise-grade autonomous AI orchestration engine implementing Model Context Protocol (MCP) to manage GitHub repositories, branches, files, issues, and pull requests via Gemini AI natural language instructions, featuring built-in human-in-the-loop security guardrails, destructive action interception, and granular audit logging.
-
Independent security verification for code written by humans and AI agents. Scan, repair, then prove it — Dvalin runs your project's own checks and issues a Verified Fix Record anyone can re-derive offline. Local-first, policy-bound, MIT.
npm MIT ★ 121 -
GoBuildWP (formerly WPPilot): WordPress MCP server & AI agent for ChatGPT, Claude, Cursor, Codex. Build and edit full websites in Elementor and Gutenberg, run SEO, PageSpeed checks, safe updates, backups and security, many sites via GoBuildWP Cloud. Safety profiles, human approval, undo. Free core; Pro adds WooCommerce and more builders.
GPL-2.0 ★ 65