Agentgate
Approval workflows for AI agents
What is Agentgate?
Agentgate is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to approval workflows for ai agents
Approval workflows for AI agents
This server falls under the Security category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Approval workflows for AI agents
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx agentgateConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Agentgate
Agentgate is a self-hosted approval workflow server for AI agents that intercepts sensitive actions and routes them to a human for approval before execution. It provides a policy engine that auto-approves safe actions, auto-denies risky ones, and routes others to Slack, Discord, email, or a web dashboard. The MCP integration exposes tools like agentgate_request_approval and agentgate_list_policies that let Claude submit requests and check approval status directly, enabling human-in-the-loop control over autonomous agent workflows.
Prerequisites
- Node.js 18 or later and pnpm installed
- Docker and Docker Compose for the recommended container deployment
- A Slack Bot Token and Signing Secret (if using Slack for approvals), or Discord Bot Token (if using Discord)
- An MCP client such as Claude Desktop
- An AGENTGATE_API_KEY generated after bootstrapping the server
Clone the repository
Clone the AgentGate repository and navigate into it.
git clone https://github.com/agentkitai/agentgate.git
cd agentgateStart the services with Docker Compose
Copy the example environment file, edit it with your Slack or Discord credentials, then start all services (server, dashboard, PostgreSQL, Redis).
cp .env.example .env
# Edit .env with your credentials:
# SLACK_BOT_TOKEN=xoxb-...
# SLACK_SIGNING_SECRET=...
# DISCORD_BOT_TOKEN=...
# DISCORD_DEFAULT_CHANNEL=...
docker-compose up -dBootstrap the server to get an API key
Run the bootstrap command to create the admin user and generate the initial API key. Save this key — you will need it for the MCP client configuration.
pnpm install
pnpm --filter @agentgate/server bootstrap
# The bootstrap output includes AGENTGATE_API_KEYConfigure your MCP client
Add Agentgate to your Claude Desktop configuration. The server exposes MCP tools at the running service endpoint.
{
"mcpServers": {
"agentgate": {
"command": "npx",
"args": ["agentgate"],
"env": {
"AGENTGATE_API_KEY": "your-api-key-here",
"PORT": "3000"
}
}
}
}Create approval policies
Use the web dashboard at http://localhost:8080 or the agentgate_create_policy MCP tool to define which actions auto-approve, auto-deny, or require human review.
# Via CLI:
agentgate config show
# Example policy via agentgate_create_policy tool:
# action: "send_email" → require approval
# action: "read_file" → auto-approve
# action: "delete_database" → auto-denyRequest and await human approval
In your agent workflow, use the agentgate_request_approval tool to submit a pending action. The tool returns a request ID you can poll with agentgate_check_request until a human approves or denies it via Slack, Discord, or the dashboard.
# Claude will call these tools:
# agentgate_request_approval: { action: 'send_email', params: { to: '[email protected]', subject: 'Report' }, urgency: 'high' }
# agentgate_check_request: { request_id: 'req_abc123' }Agentgate Examples
Client configuration
Claude Desktop configuration for Agentgate with API key and server port configured.
{
"mcpServers": {
"agentgate": {
"command": "npx",
"args": ["agentgate"],
"env": {
"AGENTGATE_API_KEY": "your-api-key-here",
"PORT": "3000",
"DATABASE_URL": "./data/agentgate.db"
}
}
}
}Prompts to try
These prompts test the approval request flow, policy management, and request listing capabilities.
- "Request approval to send an email to the team with the weekly report"
- "Check the status of approval request req_abc123"
- "List all pending approval requests"
- "Show me the current approval policies"
- "Create a policy that auto-approves file read operations but requires human review for file deletions"Troubleshooting Agentgate
Bootstrap command fails with database connection error
Ensure the PostgreSQL and Redis containers are fully started before running bootstrap. Run `docker-compose ps` to verify all services are in the 'running' state. For file-based SQLite (default non-Docker), ensure the ./data directory exists and is writable.
Slack approval messages are not being delivered
Verify SLACK_BOT_TOKEN and SLACK_SIGNING_SECRET are correctly set in your .env file. Ensure the Slack bot has been added to the target channel and has the `chat:write` OAuth scope. Check the server logs with `docker-compose logs server` for Slack API error messages.
AGENTGATE_API_KEY is invalid or not recognized
Re-run `pnpm --filter @agentgate/server bootstrap` to regenerate the key, or generate a new key via the web dashboard at http://localhost:8080. Ensure the key is correctly pasted into the MCP client config env block without any trailing spaces.
Frequently Asked Questions about Agentgate
What is Agentgate?
Agentgate is a Model Context Protocol (MCP) server that approval workflows for ai agents It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Agentgate?
Follow the installation instructions on the Agentgate GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Agentgate?
Agentgate works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Agentgate free to use?
Yes, Agentgate is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Agentgate Alternatives — Similar Security Servers
Looking for alternatives to Agentgate? Here are other popular security servers you can use with Claude, Cursor, and VS Code.
Casdoor
★ 13.6kAn open-source Agent-first Identity and Access Management (IAM) /LLM MCP & agent gateway and auth server with web UI supporting OpenClaw, MCP, OAuth, OIDC, SAML, CAS, LDAP, SCIM, WebAuthn, TOTP, MFA, Face ID, Google Workspace, Azure AD
ghidraMCP
★ 9.0kAn Model Context Protocol server that enables LLMs to autonomously reverse engineer applications by exposing Ghidra's decompilation and analysis tools. It allows AI agents to list code structures, rename methods, and analyze binaries directly through
HexStrike AI
★ 8.9kHexStrike AI MCP Agents is an advanced MCP server that lets AI agents (Claude, GPT, Copilot, etc.) autonomously run 150+ cybersecurity tools for automated pentesting, vulnerability discovery, bug bounty automation, and security research. Seamlessly b
IDA Pro MCP
★ 8.7kEnables AI-assisted reverse engineering in IDA Pro by providing tools to analyze binaries, decompile functions, manage comments, search patterns, and interact with the IDA database through natural language.
Anthropic Cybersecurity Skills
★ 6.6k754 structured cybersecurity skills for AI agents · Mapped to 5 frameworks: MITRE ATT&CK, NIST CSF 2.0, MITRE ATLAS, D3FEND & NIST AI RMF · agentskills.io standard · Works with Claude Code, GitHub Copilot, Codex CLI, Cursor, Gemini CLI & 20+ platform
Hooker
★ 5.1k🔥🔥 hooker is a Frida-based reverse engineering toolkit for Android. It offers a user-friendly CLI, universal scripts, auto hook generation, memory roaming to detect activities/services, one-click SOCKS5 proxy setup, Frida JustTrustMe, and BoringSSL u
Browse More Security MCP Servers
Explore all security servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Agentgate in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Agentgate?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.