SentinelGate
- 25 repo stars
- Sentinel-Gate/Sentinelgate
SentinelGate is an MCP server that proxies agent MCP requests with CEL policies, RBAC, and auditing.
About
SentinelGate is an open-source MCP proxy that lets developers treat agent tool access like production API access: policies, roles, and audits instead of handing every MCP server full filesystem and network reach. You position SentinelGate between your coding agent and downstream MCP servers, express rules in CEL, assign RBAC, and rely on auditing when something sensitive is invoked—ideal during Ship security work and Operate when agents touch staging or internal services. Release v2.0.0 distributes platform-specific mcpb packages with pinned file SHA256 hashes for darwin and linux amd64, which suits teams that want reproducible installs without building from source. Complexity is advanced because policy design and proxy topology take thought, but the payoff is centralized guardrails across Claude Code, Cursor, and other MCP hosts. SentinelGate is governance infrastructure for MCP, not another data or browser automation connector.
- Open-source MCP proxy for AI agent access control
- CEL policies plus RBAC for fine-grained tool and resource rules
- Audit trail oriented design for agent MCP traffic
- Shipped as mcpb binaries for darwin arm64/amd64 and linux amd64 at v2.0.0
- stdio transport compatible with standard MCP clients behind the proxy
SentinelGate by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
claude mcp add SentinelGate -- npx -y Sentinel-Gate/SentinelgateAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| repo stars | ★ 25 |
|---|---|
| Repository | Sentinel-Gate/Sentinelgate ↗ |
What it does
Put an MCP proxy in front of agent tool access so CEL policies, RBAC, and audit logs gate what Claude or Cursor can call in production-adjacent workflows.
Who is it for?
Best when you're exposing multiple MCP tools to agents in staging or production and need centralized deny/allow rules and audit.
Skip if: Hobby projects with a single benign local MCP tool and no compliance or shared-environment risk.
What you get
After deployment, MCP traffic flows through SentinelGate so only policy-allowed tools run under assigned roles with auditable records.
- Policy-gated MCP proxy layer for agent sessions
- RBAC-scoped tool access across chained MCP servers
- Audit-oriented handling of agent MCP invocations
By the numbers
- Server version 2.0.0
- 3 published mcpb platform packages (darwin arm64, darwin amd64, linux amd64)
- stdio transport
Recommended MCP Servers
1Claw VaultHSM-backed vault secrets for AI agents (JIT fetch) plus prompt-injection and threat scanning.2
1passwordMCP server for 1Password service accounts — tools and resources for vaults and credentials18
402sentinel McpAssess an x402 counterparty's risk BEFORE paying: allow/review/block, scored on-chain (Base).1
A2adependencyinspector McpRemote MCP for A2A dependency inspector MCP, structured receipts, audit logs, and reviewer-ready evi
A2a Governance Bridge McpA2A Governance Bridge MCP server.A2aidentitytoll McpRemote MCP for A2A caller identity, scope policy, verdict receipts, and audit history.How it compares
MCP access-control proxy, not an MCP server that fetches data or compiles documents.
FAQ
Who is SentinelGate for?
Developers shipping agent-assisted workflows who must limit which MCP tools run, for which roles, with an audit trail.
When should I use SentinelGate?
Use it before agents touch sensitive integrations—payments, deploy hooks, internal APIs—especially during security review and production operation.
How do I add SentinelGate to my agent?
Download the matching v2.0.0 mcpb release for your OS/architecture, configure CEL policies and RBAC, point your MCP client at the SentinelGate stdio proxy, and chain downstream servers behind it.