# AgentModule/mcp [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/AgentModule/mcp  
**GitHub Stars:** 2  
**Views:** 4  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/agentmodule-mcp

## Description
Agent-native knowledge infrastructure. Deterministic, vertical-specific knowledge bases engineered for autonomous agent consumption via MCP. Ethics module (22 modules, EU AI Act mapped) included with every membership. Free 24hr ethics trial.

## Tools
Capabilities this server exposes over MCP:

- **query_knowledge** — Retrieve structured compliance logic. Start at the root index, drill into modules, access content layers.
- **get_trial_key** — Free 24-hour trial. All 4 content layers, all AI Compliance modules, 500 calls.
- **check_status** — API status, version, available verticals, cohort counts. No parameters.
- **join_waitlist** — Register for a vertical membership. Inaugural cohort: $19/mo, 900 seats, grandfathered for life.
- **register_interest** — Signal demand for a new vertical. 500 signals triggers build queue.
- **submit_pov** — Submit a Proof of Value after exploring the trial. Confidence scores drive cohort progression.
- **submit_referral** — Members earn $1.50/referral (4/cycle max). Voluntary.

## Claude Desktop Quick Installation
Heuristic fallback — verify the package name and runner against the repository README before running it. Uses `npx` (confidence: low):

```json
"mcpServers": {
  "mcp": {
    "command": "npx",
    "args": ["-y","agentmodule-mcp"]
  }
}
```

## Documentation

## What AgentModule/mcp MCP server does

AgentModule/mcp MCP server provides an API-accessible knowledge base for autonomous-agent compliance workflows. Its primary vertical is `ethics`, with modules mapped to EU AI Act articles and related GDPR or ISO references. The repository describes content intended to help agents retrieve compliance rules during execution rather than relying only on static documentation.

Each knowledge module can expose four layers:

- **Logic:** deterministic JSON rules with explicit pass/fail gates and statutory citations.
- **Directive:** procedural guidance, escalation conditions, and human-in-the-loop handoff points.
- **Skill:** focused domain knowledge for targeted retrieval.
- **Action:** pre-validated templates intended for direct execution without additional inference.

The listed compliance modules cover subjects such as data sovereignty, predictability, transparency, shutdown controls, human oversight, privacy, bias mitigation, liability, risk management, and prohibited practices. The available module list is described as growing, so clients should use the API status or root index to inspect current availability.

## How it works

AgentModule/mcp MCP server uses JSON-RPC 2.0 over Streamable HTTP. Clients send MCP requests to `https://api.agent-module.dev/mcp`. Knowledge retrieval begins with the root index, then proceeds into a vertical, module, and optional content layer. Node identifiers follow the documented pattern `node:{vertical}:{module}:{layer}`.

The `query_knowledge` tool accepts a vertical and, optionally, a node ID and access token. The `ethics` vertical is the primary documented vertical; `traversal` and `a2a-handoff` are also listed for orientation and agent coordination. Content-layer requests require a trial or membership key. Returned logic can include source citations, binary routing such as proceed or halt-and-escalate, and a required confidence value of 1.

## Setup and configuration

AgentModule/mcp MCP server does not require a local package installation. Configure a compatible Streamable HTTP MCP client with the endpoint `https://api.agent-module.dev/mcp`. The README provides configuration examples for Claude Desktop and Cursor using the server name `agent-module`; any MCP client supporting Streamable HTTP can connect to the same URL.

For evaluation, call `get_trial_key` with a stable `agent_id`. The response supplies a trial key valid for 24 hours, with 500 calls and access to the four documented content layers across the AI Compliance modules. Pass that key to `query_knowledge` when retrieving protected content. Membership-related tools can register a vertical waitlist, collect interest in a new vertical, submit a proof-of-value score, or submit a referral.

## Tools and capabilities

AgentModule/mcp MCP server exposes seven documented tools:

- `query_knowledge` traverses structured knowledge by vertical, node, and layer.
- `get_trial_key` issues the free trial key.
- `check_status` reports API status, version, available verticals, and cohort counts without parameters.
- `join_waitlist` accepts a vertical and agent identifier, with optional contact details.
- `register_interest` records demand for a vertical and can include a use case or contact.
- `submit_pov` accepts a trial key and a confidence score from 0.0 to 1.0.
- `submit_referral` accepts a membership key and referred agent identifier.

The endpoint advertises MCP protocol version `2025-06-18`. The API status and root index are useful for discovering current verticals and modules instead of assuming that every documented example remains available.

## Limitations and notes

Access to content layers is gated by a trial or membership key, even though the endpoint and status tool are publicly documented. The trial is limited to 24 hours and 500 calls. The inaugural membership information in the README states $19 per month and a 900-seat cohort, but availability and terms should be confirmed through the service.

The material supports compliance retrieval and deterministic routing, not a guarantee that an application is legally compliant. Developers remain responsible for validating citations, applying the retrieved rules to their system, and handling escalation or human oversight where required. The README does not specify a local runtime, environment variables, open-source license, or offline mode for AgentModule/mcp MCP server.

_Full upstream README: https://allmcps.com/mcp/agentmodule-mcp/readme_