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 metadata comes from ecosyste.ms (CC BY-SA 4.0). 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.
-
Prompt-injection guardrail with a hardened LLM judge, guarded RAG and an MCP server · FastAPI + React/TS
MIT -
Guardrails for LLM answers. GroundLens checks an LLM answer against its sources without a model, decides PASS, REVIEW or FAIL under your policy, and keeps a signed record.
pypi Apache-2.0 ★ 8 -
sentinel-aigithub.com/MaxwellCalkin/sentinel-aiupdated this year
Real-time AI safety guardrails for LLM apps. 10 scanners: prompt injection, PII, harmful content, code vulnerabilities, obfuscation detection. Sub-ms latency. Python + TypeScript SDKs. MCP proxy. Claude Code hooks.
Apache-2.0 ★ 27 -
Evaluation and runtime governance layer for local LLMs (Ollama + MLX-LM): deterministic policy/provenance guardrails, MCP server, regression gate, and reviewer console
★ 1 -
Multi-agent LLM research platform: parallel data agents, synthesis agents, deterministic guardrails, MCP server
MIT ★ 1 -
A 3-stage safety guardrail agent for LLM coding assistants (Claude Desktop, Cursor, CodeX) via MCP protocol.
MIT ★ 3 -
Guardrails for LLM answers. GroundLens checks an LLM answer against its sources without a model, decides PASS, REVIEW or FAIL under your policy, and keeps a signed record.
Apache-2.0 -
Constitutional-AI safety guardrails for any LLM — MCP server (npm: @creedspace/mcp-server, registry: io.github.Creed-Space/creedspace-mcp-server).
npm MIT -
Make ordinary LLMs smarter via offline cognitive augmentation. MCP server: guardrails, self-critique, verification, plan/task management. No API keys.
MIT -
An LLM agent that fixes git mistakes from plain-language requests, safely: sandboxed runs, auto backups, human approval for destructive steps, and hard guardrails. CLI + MCP server.
MIT -
claude-code-invoice-guardgithub.com/mikko-lab/claude-code-invoice-guardupdated this year
Deterministic guardrail (refuse, don't guess) for an LLM agent on a VAT-critical invoice workflow: skill + subagent + MCP server + PreToolUse hook. Portfolio/demo plugin.
MIT -
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 -
Three-agent research system on the OpenAI Agents SDK — Planner/Executor/Reviewer with typed handoffs, a custom MCP server, function tools, input/output guardrails, token-streaming (CLI + SSE), tracing, and an LLM-judge eval harness with a pytest regression gate.
-
excellmgithub.com/mroshdy91/excellmupdated this year
ExceLLM - MCP server for Excel live automation with LLM safety guardrails
pypi MIT License -
Guardrails service for AI agents. Default-deny tool call evaluation with LLM safety analysis, priority-ordered decision matrix, and human-in-the-loop escalations. Session recording, behavioral analysis, MCP proxy, secret redaction, and real-time audit.
pypi NOASSERTION ★ 22 -
Policy-as-code for FERPA + Title IV: a deterministic guardrail that decides before any LLM speaks. Risk-tiered, citation-faithful, with an MCP server, OSCAL/SARIF exports, a GitHub Action, and a live web demo.
pypi Apache-2.0 ★ 1 -
annuma-study-companiongithub.com/AlirezaFakari/annuma-study-companionupdated this year
AI active-recall study agent for AnNuMa — MCP server, read-only guardrail, LLM-as-judge
★ 1 -
MCP server giving an LLM read-only SQL access to an ERP database, with layered guardrails
MIT -
Currency conversion agent using Google ADK, MCP, A2A with LLM Guardrails
Apache-2.0 -
LLM cost calculator, token counter, latency benchmark, CI guardrail, MCP server, and VS Code/Cursor extension.
npm MIT ★ 1 -
Safety-first MCP server connecting LLM agents to the Binance Spot API (testnet-first, read-only tools, guardrails).
-
MCP servers that let LLM agents operate Juju safely: server-side policy guardrails, audit log, Prometheus-driven autoscaling, and an eval harness for open-weight models.
Apache-2.0 -
trade-mcpgithub.com/Aaditya2502/trade-mcpupdated this year
MCP server exposing trading analytics (indicators, portfolio, risk, backtests) to LLM agents, with strict read/write separation and human-in-the-loop guardrails.
MIT -
Copro: an MCP coprocessor for Claude Code. Delegates side tasks (reading files, summaries, first-pass reviews, extraction) to cheap cloud or local LLMs with budget routing, fallback chains, quota awareness and guardrails.
MIT ★ 1 -
commandergithub.com/drhalto/commanderupdated this year
Local control for an EZ-Robot JD humanoid, replacing Synthiam ARC - web GUI, safety guardrails, and LLM agents as the robot's operator over MCP. No account, no subscription, no cloud.
MIT -
An autonomous LLM fund that scores its own stock predictions — Brier score, calibration curves, eval + ablation harness — and serves its full decision history over MCP. LangGraph pipeline with deterministic risk guardrails.
★ 1 -
AI SRE assistant: LangGraph ReAct agent + custom MCP server (Docker/Azure Monitor telemetry) + hybrid RAG (pgvector, BM25, RRF, cross-encoder) over runbooks & post-mortems, with guardrails and LLM evals
MIT -
unlimited-mcpgithub.com/triumsebas/unlimited-mcpupdated this year
Stop burning your AI limits. MCP server that delegates coding and sysops work to any cheaper agent or local LLM via durable background queues — git-worktree isolation, server-enforced guardrails, clarification rounds.
pypi Apache-2.0 ★ 2 -
@hasna/microservicesgithub.com/hasna/microservicesupdated this year
21 production-grade microservice building blocks for AI-native SaaS — auth, billing, LLM gateway, agent registry, RAG, guardrails, tracing, and more. Each with PostgreSQL, HTTP API, MCP server, and CLI.
npm Apache-2.0 -
shieldpi-mcpgithub.com/shieldpi1/shieldpi-watchtowerupdated this year
MCP server for ShieldPi Watchtower — run LLM security scans, generate drop-in guardrails from scan findings, operate the AI SOC analyst (incident triage + kill-chain reconstruction), and query the live agent monitor from Claude Desktop / Cursor / Continue.
pypi MIT -
MCP server that turns an Obsidian vault into working memory for an LLM agent. Deterministic write guardrails in code, hybrid search (BM25 + vector + temporal + graph), automated retention pruning. Built for Claude Code, works with any MCP client.
MIT -
Zero-trust AI control plane: strict stdio MCP server, MCP security gateway (JWT + RBAC), streaming PII guardrail, and a rate-limited failover LLM router, with an end-to-end demo and 348 tests.
-
MCP server for the Pionex exchange with strict anti-hallucination guardrails: two-phase order confirmation, live symbol/price validation, operator-set limits. Works with Claude and local LLMs.
MIT -
Telegram <-> Claude Code bridge: control your PC's AI agent sessions, files and terminals from Telegram, with hard security guardrails enforced outside the LLM.
MIT -
Conversational FRC telemetry analysis with built-in guardrails against LLM overconfidence. MCP server for WPILib and REV logs.
MIT ★ 14 -
OmniData-MCPgithub.com/MostafaAI10/OmniData-MCPupdated this year
Unified MCP server for data engineering & analytics lets LLMs query, profile, transform, and visualize data via DuckDB and PySpark, with read-only guardrails and local audit logging.
MIT ★ 1 -
flowproxygithub.com/dbose/flowproxyupdated this year
Self-hosted dbt-core + MetricFlow semantic layer proxy. Query governed metrics from Amazon QuickSight and LLM agents over the PostgreSQL wire protocol, with guardrails enforced. Air-gapped, no dbt Cloud, no custom drivers.
Apache-2.0 -
eu-compliance-guardgithub.com/vishtechie07/eu-compliance-guardupdated this year
A Model Context Protocol (MCP) server that enables AI agents to audit code and documentation against the EU AI Act and regional regulations. Seamlessly integrate compliance guardrails into your LLM-driven development workflow.
MIT -
LLM-nexusgithub.com/saanvijay/LLM-nexusupdated this year
A transparent HTTPS proxy between any AI client and upstream LLMs — redacts PII, compresses prompts, caches semantically similar responses, and surfaces everything in a real-time dashboard. Configurable via a single JSON file with upstream proxy chaining and MCP server integration.
MIT -
mcp-guardrailgithub.com/mini4ai4/mcp-guardrailupdated this year
An MCP server that gives an AI agent read access to a database without giving it the database — enforced by SQLite's authorizer, proven by an adversarial fuzz harness.
MIT -
injection-pii-guardrail-mcpgithub.com/ss-2303/agent-guardrail-mcpupdated this year
An MCP server providing prompt-injection detection, PII/secrets redaction, and an audit trail for AI agents.
pypi MIT -
agent-guardrailsgithub.com/logi-cmd/agent-guardrailsupdated this year
Checks AI-generated code changes before merge: scope, validation, risk, review evidence, and optional Verity receipts.
npm MIT ★ 28 -
An MCP guardrail that enforces architecture conventions while AI writes code.
MIT -
Enterprise multi-language security MCP — advanced taint, tree-sitter, plugins, SARIF/SBOM, policy gates, dashboard
MIT -
Deterministic CI/CD gate and native MCP server to block AI coding agents from slopsquatting, breaking architecture, and bypassing linters.
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 -
Streaming privacy for LLM apps: pii-leak-benchmark (does your gateway leak PII in streams?), chunk-invariance (one-line test for split-boundary leaks), mcp-ssrf-check, and a self-hosted redacting gateway.
pypi Apache-2.0 ★ 17 -
OmniDoc-QA-LLM-RAGgithub.com/aniketpoojari/OmniDoc-QA-LLM-RAGupdated this year
Agentic NL-to-SQL assistant using Model Context Protocol (MCP) with multi-layer guardrails and token cost tracking.
MIT -
llm-eval-searchgithub.com/fvahedian/llm-eval-searchupdated this year
—
-
Stop RAG hallucinations: verify every LLM answer against its sources and say "I don't know" instead of guessing. TypeScript, Python, MCP server.
npm Apache-2.0 -
promptdefend-mcpupdated this year
Prompt Defend MCP Server - 16-Layer Guardrails for AI Agents
pypi Proprietary -
Reality-check layer for AI agents. Detect hidden risks, missing context & real-world consequences before your LLM answers or acts. MCP tool for Claude Code & Cursor.
NOASSERTION -
Typed judgment tools for MCP agents. TypeSafe's Jev model as verify, screen, find, classify, rerank, decide, compare, extract, review, gate, and score: the model judges, policy decides auto, review, or escalate.
MIT ★ 49 -
metabase-mcp-progithub.com/orlik9l/mcpsupdated this year
Agent-first Metabase MCP server with a pluggable, table-scoped query-guardrail layer.
pypi MIT -
A deterministic policy firewall for AI agent tool calls. YAML rules, no LLM calls, no risk scores. CLI + MCP server (advisory) + @enforce() decorator (unbypassable). MIT, 46 tests, zero network dependency.
pypi MIT ★ 1 -
promptdefend-mcpgithub.com/promptdefend/promptdefend-mcpupdated this year
Prompt Defend MCP Server - 16-Layer Guardrails for AI Agents
npm SEE LICENSE IN LICENSE -
Tailored resumes from one fact library: page-target builds, fact guardrails, shortlist-ai scoring, CLI + MCP server
MIT -
Local MCP server and CLI for learned AI guardrails, memory, skills, and agent config injection
pypi Source Available — free for personal use, commercial/government use requires wri -
Guarded multi-turn recruiting agent: MCP server, LangChain MCP adapters, LangGraph, conversational evals with guardrail ablation, FastAPI, Docker, CI.
MIT -
Zero-dependency runtime control for production Python agents - stop loops, retry storms, and budget burn
pypi MIT