# ejentum-mcp

Supplies models with structured cognitive procedures and reasoning DAGs before generation. Built for complex, multi-constraint planning and code tasks.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 16
- Author: ejentum
- License: MIT · policy=link-only · description only; the file stays with the author
- Source: https://github.com/ejentum/ejentum-mcp
- Updated: 2026-06-11
- HTML: https://affpapa.org/en/nevibecode/mcp/ejentum-mcp

## What it does

The server connects coding assistants to a catalogue of 679 cognitive procedures. Before writing an answer, the model invokes one of eight tools based on a task prompt. It receives a structured protocol with falsification criteria, alongside a directed acyclic graph that defines decision gates, branch points, and bounded loops.

The toolset splits across two execution tiers. Standard dynamic tools (`reasoning`, `code`, `anti-deception`, and `memory`) perform semantic retrieval over abstract patterns. Adaptive variants run an intermediate model to re-label every node with project-specific identifiers. That adaptation step adds 2 to 3 seconds of latency and requires a Go or Super subscription tier.

Running the server requires an `EJENTUM_API_KEY`. You can launch it locally via stdio with Node.js 18+ or route calls to the hosted endpoint at `https://api.ejentum.com/mcp`. Keep in mind that the `memory` tool acts strictly as a perceptual filter, so calling it for factual extraction breaks its intended workflow.

## Requirements
- Needs: Node.js (commands run via npx)
- Remote server: `https://api.ejentum.com/mcp`
- Environment variables: `EJENTUM_API_KEY` — keys and tokens: ask the human for the values, never commit them or put them in shared files.

## Install

### Claude Code (native platform)

1. **Platform CLI**
```bash
claude mcp add ejentum-mcp -e EJENTUM_API_KEY=<EJENTUM_API_KEY> -- npx -y ejentum-mcp
```
Replace <…> with your own values

2. **Platform CLI: remote server**
```bash
claude mcp add --transport http ejentum-mcp https://api.ejentum.com/mcp
```

### Codex

1. **Platform CLI: remote server**
```bash
codex mcp add ejentum-mcp --url https://api.ejentum.com/mcp
```

2. **Platform CLI**
```bash
codex mcp add ejentum-mcp --env EJENTUM_API_KEY=<EJENTUM_API_KEY> -- npx -y ejentum-mcp
```
Replace <…> with your own values

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "ejentum-mcp": {
            "url": "https://api.ejentum.com/mcp"
        }
    }
}
```
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "ejentum-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "ejentum-mcp"
            ],
            "env": {
                "EJENTUM_API_KEY": "<EJENTUM_API_KEY>"
            }
        }
    }
}
```
Replace <…> with your own values

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "ejentum-mcp": {
            "type": "http",
            "url": "https://api.ejentum.com/mcp"
        }
    }
}
```
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "ejentum-mcp": {
            "type": "stdio",
            "command": "npx",
            "args": [
                "-y",
                "ejentum-mcp"
            ],
            "env": {
                "EJENTUM_API_KEY": "<EJENTUM_API_KEY>"
            }
        }
    }
}
```
Replace <…> with your own values

### Gemini CLI

1. **Platform CLI: remote server**
```bash
gemini mcp add -t http ejentum-mcp https://api.ejentum.com/mcp
```

2. **Platform CLI**
```bash
gemini mcp add -e EJENTUM_API_KEY=<EJENTUM_API_KEY> ejentum-mcp npx -y ejentum-mcp
```
Replace <…> with your own values

## When to use it
- guide complex multi-step reasoning with formal decision topologies
- counter sycophancy and hallucinations during technical evaluations
- filter prompt context through structured integrity checks
- rewrite abstract engineering patterns into task-specific DAG steps

## Example prompts
- Call the `code` tool with our database migration prompt to fetch a structured verification flow
- Run `anti-deception` before evaluating this pull request architecture
- Use `adaptive-reasoning` to map edge cases for this distributed lock logic

## Similar
- [prompt-engineering](https://affpapa.org/en/nevibecode/skills/spencerpauly-prompt-engineering.md) — skill · 413 installs
- [orchestration](https://affpapa.org/en/nevibecode/skills/orchestration.md) — skill · 350300 installs
- [thought-based-reasoning](https://affpapa.org/en/nevibecode/skills/thought-based-reasoning.md) — skill · 1400 installs
- [AgentDB Memory Patterns](https://affpapa.org/en/nevibecode/skills/agentdb-memory-patterns.md) — skill · 1100 installs
- [autonomous-loops](https://affpapa.org/en/nevibecode/skills/autonomous-loops.md) — skill · 4600 installs
- [Lotus Wisdom](https://affpapa.org/en/nevibecode/mcp/lotus-wisdom.md) — MCP server · ★ 36

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md