# tibet-pol-mcp

**MCP server for tibet-pol — machine & process health monitoring with TIBET provenance.**

Run infrastructure integrity checks via JSON templates. Every step is verified, dependency-chained, and signed with [TIBET](https://datatracker.ietf.org/doc/draft-vandemeent-tibet/) tokens. Like htop meets audit trail.

Part of the [TIBET ecosystem](https://humotica.com) by [HumoticaOS](https://github.com/Humotica).

## Install

```bash
pip install tibet-pol-mcp
```

## Claude Code / Claude Desktop Config

```json
{
  "mcpServers": {
    "pol": {
      "command": "tibet-pol-mcp",
      "env": {
        "TIBET_POL_TEMPLATES": "/path/to/your/templates"
      }
    }
  }
}
```

## Available Tools

| Tool | Description |
|------|-------------|
| `pol_check` | Run health check against a template |
| `pol_templates` | List available check templates |
| `pol_template_info` | Get template details and steps |
| `pol_diff` | Compare two runs — track infrastructure drift |
| `pol_report` | Generate HTML report from results |
| `pol_quick_health` | One-shot health percentage + failed steps |

## How It Works

Define infrastructure checks as JSON templates:

```json
{
  "process_id": "my_stack",
  "name": "My Infrastructure",
  "steps": [
    {
      "id": "api_health",
      "name": "API responds",
      "check": "curl -sf http://localhost:8000/health",
      "critical": true,
      "intent": "Core API must be reachable"
    },
    {
      "id": "db_connected",
      "name": "Database connection",
      "check": "pg_isready -h localhost",
      "dependencies": ["api_health"],
      "critical": true
    }
  ]
}
```

Then check via MCP:

```python
pol_check("my_stack")
# → 100% health, all steps valid, TIBET token chain

pol_quick_health("my_stack")
# → {"health": "100%", "checksum": "2/2", "failed_steps": []}
```

## Step Statuses

| Status | Meaning |
|--------|---------|
| `valid` | Check passed |
| `drift` | Was passing, now failing (state change) |
| `broken` | Check failed |
| `blocked` | Dependency failed, step skipped |
| `pending` | Not yet checked |

## Template Features

- **Dependencies:** Steps can depend on other steps (topological sort)
- **Critical flags:** Critical failures block dependent steps
- **Remediation:** `on_failure` commands for auto-fix
- **Intents:** Document WHY each check exists
- **Success criteria:** Min percentage, all-critical rules
- **Alerts:** Commands on critical failure or threshold breach

## Environment Variables

| Variable | Default | Description |
|----------|---------|-------------|
| `TIBET_POL_TEMPLATES` | `/srv/jtel-stack/tibet-pol-processes` | Templates directory |

## Related TIBET Packages

- [`tibet-audit`](https://pypi.org/project/tibet-audit/) — Core TIBET provenance
- [`tibet-pol`](https://pypi.org/project/tibet-pol/) — CLI/Python API (dependency)
- [`tibet-triage-mcp`](https://pypi.org/project/tibet-triage-mcp/) — Process triage MCP server
- [`tibet-phantom-mcp`](https://pypi.org/project/tibet-phantom-mcp/) — Cross-device AI sessions
- [`tibet-ipoll-mcp`](https://pypi.org/project/tibet-ipoll-mcp/) — AI-to-AI messaging

## License

MIT — HumoticaOS