gateway: MCP server (Security) in 2026
MCP gateway with runtime security policy, tool-call-level control, and audit of agent actions.
Reachable on 2026-10-27: sign-in required, tools not listed Rank 11 of 14 in Security. Source: MCP Registry record, checked 2026-10-27.
Key facts
| Type | MCP server, remote (streamable HTTP) |
|---|---|
| Category | Security |
| Endpoint | https://mcp.sealgate.ai/mcp |
| Registry name | ai.sealgate/gateway |
| Version (registry) | 1.0.0 |
| Published in registry | 2026-09-01 |
| Registry record updated | 2026-09-01 |
| Repository | https://github.com/Edison-Watch/sealgate-mcp |
| Website | https://sealgate.ai |
| Authentication | Sign-in required (the endpoint answered 401 with an authentication challenge) |
| Latest check | Reachable, sign-in required on 2026-10-27, 193 ms to initialize |
| Checks passed | 0 of 1 daily checks |
| Source | https://registry.modelcontextprotocol.io/v0/servers/ai.sealgate%2Fgateway/versions/latest (pulled 2026-10-27) |
Verification history
One probe a day: initialize, then tools/list, nothing else. Last 1 checks: 10-27 sign-in
Install
Claude Code
claude mcp add --transport http gateway https://mcp.sealgate.ai/mcpClaude (claude.ai, Desktop)
Settings > Connectors > Add custom connector. Name: gateway. URL: https://mcp.sealgate.ai/mcp
Cursor
{
"mcpServers": {
"gateway": {
"url": "https://mcp.sealgate.ai/mcp"
}
}
}Save as ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (this project).
ChatGPT
Settings > Connectors > Advanced > Developer mode, then Create. Name: gateway. MCP server URL: https://mcp.sealgate.ai/mcp. Authentication: OAuth.
Custom connectors need a plan with Developer mode; menu names change between releases.
Hermes Agent
mcp_servers:
gateway:
url: "https://mcp.sealgate.ai/mcp"Add to ~/.hermes/config.yaml, then run /reload-mcp.
Other security mcp servers
- Verity Verified working on 2026-10-27 · #1 in Security
Cited product-compliance ground truth for AI agents. Never generates; always cites. - tech-risk Verified working on 2026-10-27 · #2 in Security
MCP security scanner: tool-poisoning & prompt-injection audit (OWASP LLM01/08) + Base data - VerifyPulse Email Verification & DNS Security Verified working on 2026-10-27 · #3 in Security
Real-time RFC 5322 email verification, SPF/DMARC DNS audits & B2B leads (x402 Base USDC). - sectora Verified working on 2026-10-27 · #4 in Security
Threat intel + your scans/findings/Shield posture. CVE, EPSS, KEV, package vuln lookup, DAST. - MCP Marketplace Verified working on 2026-10-27 · #5 in Security
Search and install 4,000+ security-scanned MCP servers from inside any MCP-aware AI client.
Within a category, entries sort by (1) latest verification: verified working, then reachable but sign-in required, then failing; (2) share of daily checks that passed; (3) number of sourced facts attached to the entry; (4) median response time of passing checks; (5) name. Nothing else counts. Rankings are never for sale: sponsors and paid plans cannot change an entry's position. Markdown · JSON