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.
-
ontology-mcp-server-RL-Stable-Baselines3github.com/shark8848/ontology-mcp-server-RL-Stable-Baselines3updated this year
—
MIT ★ 3 -
Prove who can read your Postgres/Supabase rows — before an attacker does.
npm MIT -
Financial Data Model Context Protocol server - Connect AI models to financial data with FastAPI + PostgreSQL
★ 2 -
MCP server: let Cursor / Claude Code check a Supabase app for an exposed service_role key and broken RLS while it builds it
npm MIT -
rlsgithub.com/etherled/rlsactive
Local code search for AI agents: one exe, CLI for humans + MCP for AI
NOASSERTION ★ 1 -
RL & control-aware MCP memory system for structured experiment tracking, theory audit, and safe pattern reuse.
MIT -
Read-only Supabase RLS check that explains, in plain Arabic, which tables anyone can read or change. CLI + MCP server; no keys, no connection.
npm MIT -
Read-only checks for Supabase RLS, policies, SECURITY DEFINER grants and public buckets. CLI + MCP server.
pypi MIT -
A complete Model Context Protocol (MCP) Server for Ollama with **28 tools**, enabling the use of Ollama models via the MCP protocol.
NOASSERTION -
Cursor plugin: Supabase Row Level Security rules, quick-check skill and read-only MCP server
MIT -
Build an autonomous agentic AI system that detects fraud in banking transactions using reinforcement learning for hyperparameter optimization, integrated with MCP servers for tool access and data management.
★ 1 -
MCP server for Airlock — let an AI coding agent (Cursor, Claude) scan a deployed site for an exposed Supabase service_role key and audit Row Level Security, mid-build. Catches the #1 vibe-coding leak.
npm MIT -
Finds what a visitor, or another signed-in user, can read and change in a Supabase app by running its migrations offline, explains it in plain words, and tells your AI builder what to fix. CLI, GitHub Action and MCP server.
npm MIT -
playwright-mcp-rlgithub.com/karthikm15/playwright-mcp-rlupdated this year
Build a performant PPO policy that learns to complete browser form tasks via the Playwright MCP server, starting with single-field submit and scaling to multi-field forms with validation and dynamic DOM changes. Train on a small set of expert demonstrations using behavior cloning, then improve generalization to unseen pages through interaction.
-
Probe RL environment verifiers for false passes and false fails
pypi Apache-2.0 ★ 1 -
1C_MCP_DATAgithub.com/maximf71/1C_MCP_DATAupdated this year
Read-only MCP server for secure natural-language access to 1C:Enterprise data with user RLS
★ 4 -
The PocketBase experience on PostgreSQL: a PocketBase fork with horizontal scaling, RLS, realtime from logical replication, pgvector and MCP.
MIT -
mcp-supabase-latamgithub.com/sarteta/mcp-supabase-latamupdated this year
MCP server for Supabase + pgvector KB — multi-tenant (RLS), Spanish-first, LATAM SMB focus.
MIT -
Open-source multi-tenant backlog for humans and AI agents, with scoped REST and MCP access, PostgreSQL RLS, and self-hosting.
MIT ★ 2 -
Governed, read-only Postgres access for AI agents: an MCP server with AST-level SQL safety, hidden columns, RLS, budgets, and a tamper-evident audit log.
npm MIT -
schema-navigator-mcpgithub.com/ivnv00/schema-navigator-mcpupdated this year
MCP server for PostgreSQL schema introspection — tables, columns, FKs, RLS policies, constraints, functions. Works with Supabase, Neon, Railway, CockroachDB, local Postgres, Aurora.
npm MIT -
Postgres MCP server with OAuth 2.1 authentication, Keycloak/Authentik JWT verification, AST SQL guardrails, and role/RLS data isolation for Claude Code & Cursor.
MIT -
CIEL Aletheia: Sub-microsecond (1.09µs) in-memory fuel-metered verification engine for AI reasoning, MCP tool-calls, and RL rollouts (Apache 2.0)
★ 1 -
Local-first context discovery for coding agents: AST dependency closure, hybrid retrieval and a deterministic pre-flight gate before any paid LLM call. Google ADK, MCP, PostgreSQL RLS.
-
Static analysis scanner for MCP server code and multi-tenant SaaS applications. 57 deterministic rules covering tenant isolation, tool visibility, cache key scoping, RLS, IDOR, and credential vault isolation. MCP server for AI agent integration.
npm MIT -
Probe RL environment verifiers for false passes and false fails.
pypi Apache-2.0 -
asistent-MCPgithub.com/sitechfromgeorgia/asistent-MCPupdated this year
User-scoped MCP server for asistent.ge — one env var, RLS-enforced, zero config
-
mujoco-mcp-servergithub.com/Rongxuan-Zhou/mujoco-mcp-serverupdated this year
MuJoCo physics simulation as MCP tools for Claude Code — sim, render, analyze, control robots, and run RL environments.
MIT ★ 11 -
Onusgithub.com/dany701/Onusupdated this year
A sequential decision agent built for training prescription from HRV, sleep, and recovery data with offline RL and MCP tool servers.
-
mcp-office-assistantgithub.com/numoru-ia/mcp-office-assistantupdated this year
Minimal Go MCP server exposing Postgres (RLS), Gmail and Google Calendar as tools
NOASSERTION -
agentic-cortexgithub.com/albert-ying/agentic-cortexupdated this year
A personal AI operating system built on OpenClaw/Claude Code, MCP servers, and structured markdown. Feedback RL, voice cloning, Screenpipe integration.
MIT ★ 24 -
cheqlyupdated this year
Cheqly MCP server — manage workspaces, projects (checklists) and tasks via Supabase under user RLS.
npm MIT -
semantic-sql-mcpgithub.com/AshC2004/semantic-sql-mcpupdated this year
Three-microservice MCP server: SQL explorer with Postgres RLS, semantic column mapper, and role-based access control
-
Human-reviewed AI shift assistant with deterministic analytics, cited RAG, LangGraph, MCP, and PostgreSQL RLS.
MIT -
Kindredseniorcitizengithub.com/MuhammadFarid1990/Kindredseniorcitizenupdated this year
Helps seniors live safely and independently. Real-time visibility for guardians and care centers. Voice-first, AI-grounded, RLS-secured.
-
OWL 2 RL 推理引擎 + 领域适配器 + FastAPI/MCP + 本体图谱工作台:一个可快速应用本体的开源 Demo,配套《本体上手指南》
MIT ★ 4 -
Multi-agent RL pursuit engine over FastMCP: Dec-POMDP zero-sum game, Q-learning hybrids, natural-language state negotiation, Gmail API telemetry.
-
openclaw-hackathongithub.com/Trolleroof/openclaw-hackathonupdated this year
Apollo Labs RL run orchestrator: FastAPI PPO training, eval baselines, rollout GIFs, canonical RunReports, AgentMail delivery, Nia lessons, Next.js dashboard, and MCP tools.
★ 2 -
extremisgithub.com/ashwanijha04/extremisupdated this year
Persistent, learning memory for AI agents — RL-scored retrieval, intelligent forgetting, MCP-native. Works with Claude Code, Cursor, OpenAI.
MIT ★ 1 -
long-term-memorygithub.com/ihwooMil/long-term-memoryupdated this year
Intelligent memory MCP server for AI assistants — auto-save, semantic search, RL-powered retrieval, and knowledge graphs
pypi MIT -
Secure SQL proxy for AI agents — AST safety, per-agent RLS, audit log, MCP-native.
pypi MIT ★ 1 -
convotreegithub.com/salon64/convotreeupdated this year
An MCP fold tool: agents run sub-tasks in isolated branches and get back clean summaries, so the main context never fills up. Context folding, no RL.
MIT ★ 1 -
MCP stdio server exposing structured dev tools (Playwright e2e, Supabase RLS tester, tsc, test-output parsers, Next.js build) to coding agents
MIT -
Demo MCP server for a small online store: orders, products, stock and refunds with human approval, OAuth and Postgres RLS. Demo data only.
-
tuneforgegithub.com/JustVugg/tuneforgeupdated this year
An MCP server for dataset generation, fine-tuning, RL, and evaluation of LLMs — directly from your coding agent.
MIT ★ 17 -
Agentic project-management MCP server — CPM scheduling, resource optimization, risk tracking — on Supabase with RLS. Generated from a technical architecture by an automated software factory.
-
mmcp-coregithub.com/ragavrida/mmcpupdated this year
MMCP — Multi-Model Collaboration Pipeline. Auto-routes tasks to Claude, GPT, Gemini, DeepSeek with RL routing, cost optimization, and MCP server support.
pypi MIT -
a knowledge-graph and RAG system that extracts requirements into neo4j, queryable through an ollama and MCP chat client with RL tuned retrieval strategies.
-
The minimalist midnight cinema & anime hub. Built with Next.js 15, React 19, Supabase RLS, Google Gemini AI, and an in-app MCP server.
MIT -
MCP server with read-mostly, RLS-aware Supabase/Postgres tools for Claude. Tested against a local Supabase stack and plain Postgres. Personal open-source project.
MIT -
Gymnasium-style RL framework for LLM agent training — MDP environments, three-layer process reward & SFT/DPO/GRPO policy optimization. CLI + MCP ready.
★ 4 -
Hexagonal Python 3.12 commercial order pricing, exact Fraction margin-approval engine & multi-tenant security reference (PostgreSQL 17 FORCE RLS, FastAPI, MCP tools)
MIT ★ 1 -
soveragithub.com/Quantumboxai/soveraupdated this year
Open, EU-resident, all-Azure BaaS. Postgres + RLS, realtime, pgvector + Azure OpenAI, blob storage, MCP server for AI agents. The Supabase alternative built for HDS/GDPR.
MIT -
鼎江智绩 2.0 — 可追溯的企业咨询交付系统。NestJS + Next.js + PostgreSQL (FORCE RLS) + 自研 MCP Server;受治理写入、幂等命令内核、五层集成测试、架构守卫与可复现性能基准。Apache-2.0。
Apache-2.0 -
telegram-ai-agentgithub.com/abailey81/telegram-ai-agentupdated this year
Production-grade conversational AI platform — 13 intelligence engines, 6 deep learning models, Thompson sampling RL, 4-backend voice synthesis, and autonomous model improvement. 90+ MCP tools, 80+ REST endpoints.
Apache-2.0 ★ 3 -
Parlaygithub.com/sh4shv4t/Parlayupdated this year
Parlay — RL negotiation environment where LLM agents learn game theory, theory-of-mind, and strategic bluffing through self-play. OpenEnv compliant · MCP ready · GRPO-trained on Qwen2.5.
★ 1 -
Robot data infrastructure and runtime: turn video, state, actions, and telemetry into training episodes, RL environments, policy rollouts, and replayable evidence.
pypi MIT ★ 13 -
Local MCP security reviewer for AI coding agents — catches NEXT_PUBLIC_ secret leaks, missing RLS, unauth server actions, and verbose error leaks
npm MIT -
PokéDex con IA: agente con tool calling escrito a mano, servidor MCP, 3 LLMs con fallback (Groq → Gemini → Anthropic), evals y visión. Next.js + Supabase (RLS) + FastAPI.
MIT -
ai-agentic-erp-buildergithub.com/phong28zk/ai-agentic-erp-builderupdated this year
Multi-tenant AI Orchestrator platform for SMEs — domain agents (Finance, HR, Sales, Supply, Compliance), multi-channel I/O, RLS-isolated Postgres, pluggable ERP connectors. Architecture showcase.
NOASSERTION