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.
-
sty-xguard-agent-safety-mcpno update data
—
pypi -
agent-safety-mcpgithub.com/VoynichLabs/agent-safety-mcpupdated this year
Agent Safety MCP: STD/AIDS prevention + Chlorpromazine anti-hallucinogen for AI agents
MIT -
건설현장의 법정 안전문서 작성을 더 빠르고 정확하게. 산안법·기준규칙·중처법·KOSHA Guide를 기반으로 안전관리자와 현장소장의 문서 작성과 검토를 돕는 오픈소스 도구입니다.
npm MIT -
agent-safety-mcpgithub.com/LuciferForge/agent-safety-mcpupdated this year
MCP server wrapping ai-cost-guard, ai-injection-guard, and ai-decision-tracer into a single safety layer.
pypi MIT ★ 1 -
safety-agent-mcpupdated this year
MCP server for Superagent.sh API integration - security guardrails, PII redaction, and claim verification
npm MIT -
guardian-safety-mcpgithub.com/guardian-safety-gate/guardian-mcp-serverupdated this year
MCP server for Guardian Safety Gate — allow AI agents to evaluate and execute governed actions through a safety policy layer.
npm MIT -
Deterministic pre-execution safety certification for autonomous AI agents.
pypi -
MCP server exposing FDA drug-safety data (labels, FAERS adverse events, recalls, label-documented interactions) with structured outputs, plus a LangGraph agent that discovers and uses its tools. Works with Claude Desktop, Cursor and any MCP client.
MIT -
ai-safety-mcp-servergithub.com/Jack0319/ai-safety-mcp-serverupdated this year
A modular AI Safety MCP server providing safety evaluations, interpretability tools, and alignment-aware research utilities for agentic systems
Apache-2.0 ★ 1 -
onchain-safety-mcpgithub.com/aitools420/onchain-safety-mcpupdated this year
MCP server: multi-chain token-safety checker (rug/honeypot/LP-burn verdicts) for AI agents — PulseChain, Monad, Base, BSC
npm MIT -
aviation-safety-risk-mcpgithub.com/apifyforge/aviation-safety-risk-mcpupdated this year
Aviation safety intelligence for AI agents — this MCP server queries 8 live government data sources in parallel and returns a composite Aviation Safety Score (0-100) covering airline safety defects, airworthiness directives, route weather risk, sanctions exposure, and maintenance compliance.
MIT -
influencer-brand-safety-intelligence-mcpgithub.com/apifyforge/influencer-brand-safety-intelligence-mcpupdated this year
Influencer brand safety screening for AI agents and brand teams — this MCP server delivers automated creator vetting, controversy risk analysis, audience authenticity scoring, and sanctions screening through 8 callable tools that orchestrate 9 parallel data sources.
MIT ★ 1 -
Context compaction and safety gating for AI agents via TypeSafe Jev: keeps messages verbatim, no summarization. OpenAI, Anthropic, LangChain, CLI, MCP.
MIT ★ 2 -
agentmesh-mcp-servergithub.com/microsoft/agent-governance-toolkitupdated this year
MCP Server for Claude Desktop - Agent OS kernel primitives including code safety verification, CMVK multi-model review, and IATP trust
pypi MIT -
Safety-first clinical decision-support research agent: auditable state machine, Bayesian differential ranking, 300 deterministic medication-safety modules, FHIR R4, durable audit log and a dual-era MCP server for Claude, Cursor and VS Code. Research use only, not for clinical deployment.
NOASSERTION ★ 4 -
A safety layer for AI agents that spend money: payment gate, signed-approval verifier, MCP server and evals.
MIT -
neuroversegithub.com/joshua400/neuroverseupdated this year
Multilingual intelligence + memory + safety MCP server for autonomous AI agents
npm Apache-2.0 -
preflightai-mcpgithub.com/hackla-2026/nexusupdated this year
PreflightAI MCP server — pre-write context safety for AI coding agents.
npm MIT -
@attesso/mcpgithub.com/attesso/attessoupdated this year
Attesso MCP server. Authorization and safety tools for AI agents.
npm MIT -
sh-guard-mcpgithub.com/aryanbhosale/sh-guardupdated this year
MCP server for sh-guard: semantic shell command safety classifier for AI coding agents (Cursor, Cline, Windsurf)
cargo GPL-3.0-only -
@entropy0/mcpgithub.com/entropy0dev/sdkupdated this year
Entropy0 MCP server — source trust and URL safety tools for AI agents
npm MIT -
Agent Hotline MCP server — gives AI agents a place to report safety incidents they discover (hidden exploits, planted instructions, covert agent channels).
npm MIT -
opslens-mcpgithub.com/vikasrkarjigi/opslens-mcpupdated this year
Multi-agent industrial RCA with visible MCP safety gateway. Built on Anthropic Model Context Protocol.
-
Verifiable multi-agent voice-operations control plane with safety gates, MCP tools, a 2,048-case adversarial evaluation, and CI.
-
throne-mcp-gatewaygithub.com/mdfifty50-boop/throne-mcp-gatewayupdated this year
20 Production MCP Servers for AI Agent Infrastructure — Safety, Observability, Compliance & Intelligence
★ 1 -
contextlockgithub.com/LutaElbert/contextlockupdated this year
Local MCP safety layer that blocks secrets before AI coding agents see them.
npm Apache-2.0 -
aegis-defiupdated this year
A safety layer for autonomous DeFi agents - MCP server + smart contracts that simulate, scan, and protect on-chain transactions before execution.
npm MIT -
Governed AI agents with an MCP server: least-privilege tool grants, human approval for writes, PII redaction, audit log, rollback, and safety evals in CI.
MIT ★ 1 -
Local, safety-governed computer use for AI agents on Windows, with MCP, file and terminal tools, and a portable desktop app.
Apache-2.0 -
coach-kernelgithub.com/LeesungDang/coach-kernelupdated this year
Deterministic safety kernel for AI-agent training plans, with MCP data boundaries, Claude Agent SDK integration, replay evals, and fuzzing.
MIT -
ayirugithub.com/ruth411/ayiruupdated this year
Verified, machine-readable safety verdicts for AI agents running shell commands. Append-only audit log, six verification levels, deterministic risk engine. MIT.
★ 2 -
x402 pay-per-call Solana tools for AI agents: token safety, tx-failure explainer, preflight, wallet activity, pump.fun launches (USDC)
MIT -
outlook-agentgithub.com/johnkil/outlook-agentupdated this year
Local safety-gated MCP bridge for AI agents and Outlook mail & calendar: metadata-first reads, explicit content, reviewed sends, guarded cleanup.
Apache-2.0 ★ 3 -
Model Context Protocol (MCP) server for Aletheia: Sub-millisecond runtime safety & scope creep (S3) filter for AI agent tool calls
npm NOASSERTION ★ 12 -
Deterministic pre-execution safety gate for AI agents: verify tool calls against trusted evidence, policies, and WorldState before side effects.
★ 1 -
pharma-signal-mcpgithub.com/niteowlpt/pharma-signal-apiupdated this year
Drug safety intelligence MCP server. 358 drugs, 1M+ FDA adverse events, pharmacovigilance signal detection, company risk analysis. 19 tools for AI agents.
npm MIT -
agent-tripwiregithub.com/ipezygj/agent-guardupdated this year
Safety checks an AI agent runs before it acts: is this package malware/typosquat, is this shell command destructive, does this text leak a secret. CLI + MCP server.
pypi MIT -
binevalgithub.com/darksolitaire9-hub/binevalupdated this year
A deterministic, Rust-native binary evaluation kernel for AI safety, designed to help autonomous agents self-correct and red-team LLM infrastructure without fuzzy scores.
MIT -
repo-seatbeltgithub.com/berkcangumusisik/repo-seatbeltupdated this year
A safety layer for AI coding agents. CLAUDE.md/AGENTS.md generator, MCP runtime guardrail, pre-commit hook, GitHub Action.
npm MIT ★ 8 -
mcp-server-sentinelseedgithub.com/sentinel-seed/mcp-server-sentinelseedupdated this year
MCP server for Sentinel AI safety seeds - add alignment guardrails to any AI agent
npm MIT -
user-query-guardgithub.com/gowthams05/user-query-guardupdated this year
A safety gateway for validating user queries before they reach LLMs, agents, tools, or RAG pipelines.
pypi MIT -
sicarius-guardgithub.com/Chronolapse411/sicarius-guardupdated this year
Solana Token Safety Oracle for AI Agents & Trading Bots — byte-level SPL analysis, honeypot detection, Birdeye enrichment, MCP server
npm MIT ★ 2 -
Safety layer for AI coding agents: blocks dangerous commands and scripts, savepoints and rollback for project files, HMAC-signed tamper-evident audit log. MCP server + CLI.
NOASSERTION ★ 1 -
Rollback and safety for AI coding agents and vibe coders: flag risky changes and restore project files, including uncommitted work, to their pre-session state.
npm Apache-2.0 -
fhir-factguardgithub.com/gedp/fhir-factguardupdated this year
FHIR FactGuard: Zero-Trust clinical safety layer for AI agents. Validates medication claims against FHIR patient records + RxNorm. Built for Agents Assemble Hackathon 2026.
-
token-xraygithub.com/duolaamengweb3/token-xrayupdated this year
Agent-native token safety: contract + deployer reputation + LP-lock reality + bridge-in detection. stdio MCP, MIT.
npm MIT -
AI copilot that answers plain-English questions about ERP sales data. Read-only by design: LangGraph agent, MCP tools, 4-layer safety. FastAPI + SQLite + free OpenRouter models.
-
AI agent safety layer for crypto transactions - blocks honeypots, blacklists, and rug pulls before they happen
pypi MIT ★ 9 -
clinical-documentation-verificationgithub.com/kogunlowo123/clinical-documentation-verificationactive
Agentic clinical documentation verification: transcript extraction, FHIR chart reconciliation, order safety checks, consistency scoring and clinician-gated FHIR write-back (MCP, A2A, Bedrock)
Apache-2.0 -
Lightweight AI safety middleware that protects humans by intercepting self-harm and criminal intent in LLM prompts. Features a 3-stage safety pipeline, MCP server for agents, and automated care responses.
pypi Apache-2.0 ★ 27 -
deskclawgithub.com/Lelevinson/deskclawupdated this year
An AI customer-service agent that runs a real (mock) shop over WhatsApp — with hard safety rails. Skills → typed MCP tools → shared data, on OpenClaw.
MIT -
Pay-per-call pre-sign transaction safety for AI agents: EVM calldata risk -> GO/HOLD/BLOCK, settled in USDC via x402. Also an MCP server.
npm -
invariantevalgithub.com/AlpharomeroJL/invariantevalupdated this year
Declared safety-invariant eval gate: enforces your locked-field rules on every output and refuses to finalize when a model auto-fills a human-owned field. MCP tool and agent skill included.
MIT ★ 1 -
@surfaceapi/mcpactive
Surface MCP server (stdio) — the link-safety layer between an AI agent and the open web. Bridges Claude Desktop, Cursor and other stdio MCP clients to the hosted Surface MCP endpoint.
npm MIT -
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 -
Production automation for a one-person ice cream company: two PreToolUse safety hooks with 97 tests, an MCP server for traffic-aware delivery routing, and the agent layer behind Shopify, Xero and Square.
-
Asimov's Three Laws compiled to real Cedar authorization policy that acts as the runtime safety layer for AI agents which touch the physical world. Governs the real SO-101 (LeRobot) arm over MCP.
★ 1 -
Static, deterministic memory-safety checker for C/C++ that verifies code against declared contracts, with an MCP server for AI coding agents
Apache-2.0 ★ 1 -
Text2SQL data-analysis agent: three handwritten LangGraph orchestrations (deterministic pipeline / autonomous ReAct loop / Supervisor multi-agent with LLM-as-Judge), 4-layer SQL safety, MCP server+client, human-in-the-loop write path with transactional dry-run. 60-case gold benchmark.
MIT -
Multi-agent ops platform where every state-changing action is human-approval-gated by construction: LangChain 1.4 middleware, a real MCP server enforcing the same risk policy, MLflow tracing, and a CI gate that treats safety violations as absolute.
MIT