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. Each candidate is checked by a language model from its name, description and topics, and only actual MCP servers are listed (not clients, SDKs, lists or skills). Package descriptions, licences and links are read from each package's own registry (npm, PyPI, crates.io, NuGet, RubyGems; Go and Maven packages are only checked to exist), and packages the registry no longer has are removed. 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.
-
MCP-first prompt governance, refinement, and engineering evidence correlation
MIT -
claude-prompt-optimizer-mcpgithub.com/rishiatlan/prompt-optimizer-mcpupdated this year
The control plane for AI prompts — deterministic scoring, policy enforcement, audit & governance via MCP. Free tier included.
npm Elastic-2.0 -
AI Safety HOT: daily AI safety news, papers and incidents in one place — jailbreaks, prompt injection, alignment, evals, governance. Website, daily digest, RSS and MCP.
NOASSERTION ★ 2,131 -
governedmemorygithub.com/Metaworkers-ai/governedmemoryupdated this year
Open-source governed memory for AI agents: prompt-injection-resistant writes, purpose-bound retrieval, provenance, and tamper-evident audit.
pypi MIT ★ 10 -
Governed observability + governance for on-endpoint local LLMs (Ollama): model policy, deterministic prompt scanner, route-through guarding — the complement to IGEL AI Armor. 18 MCP tools (preview)
pypi MIT -
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.
-
claude-harnessgithub.com/anotherben/claude-harnessupdated this year
Governance harness for Claude Code & Codex CLI — 57 MCP tools, 45 quality-gate hooks, 64 enterprise skills, HMAC-signed evidence, 2FA infrastructure locks. Vibe prompt in, enterprise quality out.
★ 7 -
llm-guard-gatewaygithub.com/ShreyAC962/llm-guard-gatewayupdated this year
Governance gateway for LLM and agent traffic: blocks prompt injection, redacts PII, enforces per-tenant token budgets, and serves repeat prompts from a semantic cache, cutting served p50 from 44.2 ms to 4.6 ms in the included benchmark. REST + MCP surfaces, offline-first local profile.
MIT -
codex-harness-mcpgithub.com/chapzin/codex-harness-mcpupdated this year
Local Codex MCP harness: contracts, persistent RAG memory, raw traces, verification records, governance policy and PASS/FLAG/BLOCK audits, observability reports, harness profiles, eval runs, Meta-Harness-lite promotion records, natural-language harness specs, MCP resources/prompts, multi-client installer, and completion gates.
MIT ★ 9 -
Subscribe to an expert, not a prompt — a marketplace of protected expert avatar Agents served over MCP, with an anti-distillation shield, provenance, KYC, enterprise governance, and a community forum.
NOASSERTION -
@kvmfleet/mcpgithub.com/kvmfleet/mcpupdated this year
MCP server for KVM Fleet — let an AI assistant read your fleet, audit log, and compliance state and (opt-in) act on it. Includes curated governance prompts. All calls go through the platform REST API, so RBAC, policy, JIT access, and audit apply.
npm MIT -
withvibe-votergithub.com/withvibe/withvibe-voterupdated this year
Team-vote governance gate for WithVibe, as a plugin. The AI doesn't apply a prompt straight away — it opens a proposal the team votes on, and only an approved one gets built. Content-agnostic and agent-cooperative via MCP; no changes to WithVibe core.
Apache-2.0 -
MCP server for managing and sharing curated prompts across development teams
npm ISC -
Servidor MCP que expone la biblioteca de prompts AI-SDLC Pro (112 prompts, ES/EN) a agentes de IA vía stdio.
npm MIT AND CC-BY-4.0 -
🔌 Exposes AI agent prompt injection firewall over MCP — OWASP LLM Top 10 1 runtime guard. MIT.
pypi MIT -
The control plane for AI prompts. Score, enforce policy, lock config, audit prompts, and route across 4 providers / 21 costed models before any LLM call.
npm NOASSERTION ★ 2 -
Governed MCP + Skills agent platform over a legacy engineering-ops system: entitlements, audit, human-in-the-loop writes, eval-gated releases, OpenTelemetry tracing.
MIT -
🔌 Exposes OWASP Top 10 for Agentic AI security over MCP — prompt injection detection, tool poisoning, excessive agency, data leakage assessment
pypi MIT -
agent-immunegithub.com/denial-web/agent-immuneupdated this year
Adaptive threat intelligence for AI agent security — semantic memory, multi-turn escalation, output scanning, rate limiting, and prompt hardening.
pypi Apache-2.0 -
A governance layer for agent memory: write gate, read gate, verifiable forgetting, hash-chained audit trail. Sits in front of the memory store you already run.
MIT ★ 1 -
dlx7-mcp-interceptorgithub.com/everybitcounts/dlx7shieldnetupdated this year
DLX7 Agent Zero Trust (AZT) MCP server — governance runtime for AI coding agents. Validates every file edit and shell command before execution. Works in VS Code, Cursor, Windsurf, and Claude Desktop.
npm MIT -
AI Hub agent over FHIR R4 where limits live in the tools, not the prompt (with audit). Same toolset for the agent and for external MCP clients
MIT -
Not a demo — a published MCP server in the official registry. pip install guardrails-mcp. Tracks prompt regressions across labeled cases and hands a human the evidence to decide. No API key for the demo tools.
pypi MIT -
Agent-operable expense-approval console built on WebMCP. A ChatGPT browser agent triages an approval queue under server-enforced policy, with role-scoped tools, server-authorized money actions, and prompt-injection containment proven by tests.
MIT -
@cybrium-ai/mcp-servergithub.com/cybrium-ai/mcp-serverupdated this year
MCP server for AI security + AI governance + AI inventory + code security. Lets Claude, Cursor, Windsurf, Cline invoke cyscan (SAST/SCA/secrets — 1,815 rules / 75+ languages), cyweb (web vulnerability scanner), cyprobe (network discovery), cyradar (discov
npm Apache-2.0 -
veladon-mcp-pii-redactorgithub.com/grindworks-studio/veladon-mcpupdated this year
MCP server that redacts PII/PHI/secrets from outbound prompts before they leave your AI agent. Deterministic, local, auditable — built for CISOs preparing for EU AI Act Article 26 (Aug 2, 2026) and ISO 42001.
npm Apache-2.0 -
Governed Postgres access for AI assistants: an MCP gateway with OAuth 2.1 and OIDC login, one Postgres role per person, database-enforced read-only queries, PII masking, human-approved writes, a tamper-evident audit log, a red-team suite and Terraform for Cloud Run.
Apache-2.0 ★ 23 -
Connect AI coding tools to Seclai via Model Context Protocol (MCP) — manage agents, knowledge bases, and content sources from Claude, Cursor, Windsurf, and more
MIT ★ 1 -
@autodevjapan/godd-mcp-alphaactive
Governed AI development workflows that turn natural-language intent into specifications, tested code, reviews, documentation, and shippable pull requests.
npm UNLICENSED -
Architecture standards as an enforceable service: OpenAPI, data model, backend, infra and UI patterns served as MCP tools/resources/prompts, an HTTP API, and a CLI. Ships defaults; every rule is overridable per organisation.
npm MIT -
Provider-neutral control plane for running LLM agents inside an enterprise: identity, governance, knowledge, model routing and runtime. Hash-chained audit, hybrid RAG, MCP gateway, deterministic eval gate.
Apache-2.0 ★ 1 -
Fraud and scam detection for AI-agent job boards and bounties: flags fake engagement, unpaid-work risk, upfront fees, prompt injection and bait written for agents. Open audit standard (AMS-001..008), Python CLI and MCP server.
mcpb MIT ★ 1 -
Claude Code plugin marketplace for Tileward: a hook that refuses off-policy prompts before the model sees them, and a hosted MCP server that recalls only the relevant slice of a long conversation.
MIT -
A memory graph designed for the agent that uses it, not the human who feeds it. Engrama reconstructs context from associations on demand, replacing the "stuff everything into the prompt" reflex with targeted graph traversal. SQLite default, Neo4j optional.
pypi Apache-2.0 ★ 9 -
Take back control of your company's AI. Open-source private AI platform in your own cloud account: one chat for the organization, an OpenAI-compatible gateway that enforces where every prompt may go, and a library of skills, tools and context that stays yours.
Apache-2.0 ★ 1 -
元构·智能生命体(MetaGO Intelligent Lifeform)MCP Server — 55 tools (22 思维工具 + 32 核心技能 + 1 事件上报) + 8 prompts。Engine V2.1.3 硬驱动:KMWI 四层记忆 + 元进化五阶段 + 技能智能生成调用真实代码 + 927 算法驱动(via @metago-ai/algorithms)。任何 MCP 协议客户端(Claude Desktop/Cursor/Trae/ZCode 等)即开即用。基于 Model Con
npm MIT -
AgenticStore: The secure toolkit for AI agents. Instantly equip Claude Desktop, Cursor, and Windsurf with 75+ MCP tools, persistent memory, and SearXNG search, all protected by a built-in PII prompt firewall to protect your data from being exposed to AI agents.
pypi MIT ★ 5