- emisar @emisar/emisarLet AI operate servers without SSH. Choose actions, approve risky changes, and audit every step.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.337 GitHub stars.
- Lobu @lobu-ai/lobuShared, permission-aware company context for AI agents, with provenance, approvals and audit.7% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.225 GitHub stars.
- runx @runx/runxThe governed runtime for agent skills. Search the catalog and inspect a skill before running it.4% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.95 GitHub stars.
- Hexis @bevel-software/hexisGit-backed platform for skills, tools, and context for AI agents5% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.94 GitHub stars.
- arifOS — Constitutional AI Kernel @ariffazil/arifosConstitutional AI kernel with 13 MCP tools, 888_JUDGE verdict pipeline, and VAULT999 ledger.11% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.53 GitHub stars.
- Computer-Use Agents API @hcompai/hai-agentsRun and manage H Company's Computer-Use Agents from any MCP client.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.42 GitHub stars.
- Agoragentic Agent OS MCP @rhein1/agoragenticTriptych OS (Agent OS) MCP for governed routing, receipts, and USDC settlement on Base.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.40 GitHub stars.
- HealthClaw Guardrails @aks129/healthclaw-guardrailsGuardrailed FHIR access for AI agents: PHI redaction, audit trail, step-up auth, tenant isolation7% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; it needs your API key.31 GitHub stars.
- Gnt @gnt-ai/gntGit-native policy layer for AI agents: check_action verdicts against rules approved via PR.7% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; it needs your API key.30 GitHub stars.
- NEUS MCP @neus/neus-mcpThe trust harness for AI agents. Set what an agent can do before it acts.10% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.23 GitHub stars.
- IdentArk Gateway @identark/gatewayZero-secret MCP gateway for AI agents: risk-scored, audited calls with human-in-the-loop approval.10% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; it needs your API key.16 GitHub stars.
- TAP @human/tapCredential isolation for AI agents: placeholder secrets, policy checks, optional human approval.8% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.12 GitHub stars.
- Cortex Gateway @wellknownmcp/cortex-gatewaySelf-hosted federated MCP gateway: one OAuth 2.1 MCP server in front of N apps, user-level scopes.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.12 GitHub stars.
- SatGate – Economic Firewall for AI Agents @satgate-io/satgateBudget, payment and audit controls for AI agent API calls and MCP tool calls.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; it needs your API key.11 GitHub stars.
- SentinelX MCP @sentinelx/sentinelxOperate Linux, macOS and Windows from your LLM. Every action runs through an auditable allowlist.14% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.11 GitHub stars.
- Code Auditor @benahammond/code-auditor-mcpEnforces architectural invariants and code quality rules inside your AI agent's edit loop.3% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.8 GitHub stars.
- PlanWright @planwright/planwrightControl plane for autonomous software labor. Agents claim objectives over MCP with audit trail.7% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.7 GitHub stars.
- PreClick — An MCP-native URL preflight scanning service for autonomous agents (formerly URLCheck). @urlcheck/urlcheck-mcpPreClick scans links for threats and confirms intent match with high accuracy before agents click.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.6 GitHub stars.
- Axis @useaxis/axisCoding agents from Claude Code, Cursor and Codex claim jobs and lock files on one shared board.7% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.6 GitHub stars.
- Trust Oracle @new1direction/trust-oracleIndependent A-F trust grade for any MCP server, watched for drift. Free, never for sale.12% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.6 GitHub stars.
- MCP Gateway @panossalt/mcp-gatewayMulti-tenant MCP platform with OAuth 2.1, Entra SSO, RBAC and audit logging.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; no sign-in.6 GitHub stars.
AI agent governance MCP servers
A server here enforces policy on, approves, audits or restricts what AI agents are allowed to do.
Not here: servers that are merely used by agents.
733 remote1,124 localTagged by Odel's AI, which can make mistakes
Most popular in AI agent governance
By GitHub stars, at most two per publisher- Number 1: Agent-Native Dispatch@builderio/agent-native-dispatchCentral Slack/Telegram router with jobs, memory, approvals, and A2A delegation5% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.7,121 GitHub stars.
- Number 2: Omnara@omnara/omnaraDeploy durable AI agents with Omnara, an open-source managed-agent platform.6% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.2,905 GitHub stars.
- Number 3: decionis.ai MCP@decionis/mcpAuthorize consequential AI agent actions before execution5% chance this is malicious. Assessed by AI, which can make mistakes.A remote server; you sign in with your account.687 GitHub stars.
