# MetaHumotonic

> Semantic Web server platform. Knowledge Graph + MCP + AI Agent Skills.

MetaHumotonic is a self-hosted infrastructure platform that combines:
- **Neo4j Knowledge Graph** — 548K+ nodes, 535K+ relationships, ontology-first design
- **Model Context Protocol (MCP)** — connect AI agents (Claude, GPT) directly to databases
- **18 AI Agent Skills** — methodology-driven tools for development, verification, research
- **Multi-service stack** — PostgreSQL, Redis, MongoDB, MinIO, Kafka, n8n, Traefik

## Docs

- [Skill Registry](https://metahumotonic.com/skills/): 18 fetchable AI agent skills (JSON index + Markdown files)
- [MCP Gateway](https://metahumotonic.com/mcp/): Model Context Protocol endpoint for AI agent integration
- [Neo4j Browser](https://metahumotonic.com/neo4j/): Knowledge Graph explorer (548K+ nodes)
- [n8n Workflows](https://metahumotonic.com/n8n/): Visual workflow automation

## Skills

All skills are fetchable as Markdown at `/skills/{name}/SKILL.md`:

### Methodology
- [/apt](https://metahumotonic.com/skills/apt/SKILL.md): APT v21 development methodology — Anti-Rubber-Stamp, 2-Tier Taliban, Lean verification
- [/apt-sa](https://metahumotonic.com/skills/apt-sa/SKILL.md): SemanticAnchor phase — project identity bootstrap
- [/apt-sp](https://metahumotonic.com/skills/apt-sp/SKILL.md): SemanticPyramid phase — recursive decomposition
- [/apt-st](https://metahumotonic.com/skills/apt-st/SKILL.md): SemanticTwin phase — typed contract crystallization
- [/apt-scw](https://metahumotonic.com/skills/apt-scw/SKILL.md): SourceCodeWorld phase — TDD implementation

### Verification
- [/taliban](https://metahumotonic.com/skills/taliban/SKILL.md): Adversarial validation — 9-lens quality gate, GAN principle
- [/88-taliban](https://metahumotonic.com/skills/88-taliban/SKILL.md): 113-lens mathematical meta-verification (algebra, order, function, probability, complexity)

### Research & Problem Solving
- [/prometheus](https://metahumotonic.com/skills/prometheus/SKILL.md): Research-first methodology — 7-step cycle with parallel subagent research
- [/solve](https://metahumotonic.com/skills/solve/SKILL.md): Systematic problem resolution with KG recording
- [/longinus](https://metahumotonic.com/skills/longinus/SKILL.md): KG-to-source-code traceability (sourceId + sourcePath binding)
- [/harness](https://metahumotonic.com/skills/harness/SKILL.md): Architecture-as-constraint design philosophy

### Operations
- [/db-query](https://metahumotonic.com/skills/db-query/SKILL.md): Execute queries on Neo4j, PostgreSQL, MongoDB, Redis
- [/server-status](https://metahumotonic.com/skills/server-status/SKILL.md): Full server health check
- [/docker-logs](https://metahumotonic.com/skills/docker-logs/SKILL.md): Container log inspection
- [/kafka-manage](https://metahumotonic.com/skills/kafka-manage/SKILL.md): Kafka topic management
- [/deploy](https://metahumotonic.com/skills/deploy/SKILL.md): Service deployment
- [/backup](https://metahumotonic.com/skills/backup/SKILL.md): Backup management

## Infrastructure

| Service | Purpose | Protocol |
|---------|---------|----------|
| Neo4j | Knowledge Graph (548K+ nodes) | Bolt, HTTP |
| PostgreSQL | Relational database | TCP 5432 |
| Redis | Cache + pub/sub | TCP 6379 |
| MongoDB | Document store | TCP 27017 |
| MinIO | S3-compatible object storage | HTTP |
| Kafka | Event streaming (KRaft mode) | TCP 29092 |
| n8n | Workflow automation | HTTP |
| Traefik | Reverse proxy + TLS | HTTP/HTTPS |

## How AI Agents Can Use This

1. **Fetch `/skills/` index** to discover available agent skills
2. **Fetch `/skills/{name}/SKILL.md`** to load a specific skill's instructions
3. **Connect via MCP** at `/mcp/` for direct database access (Neo4j, PostgreSQL, Redis, MongoDB)
4. **Read this file (`/llms.txt`)** for a complete platform overview