Ontopix Claude Code Setup

Why This Exists

Ontopix curates a set of Claude Code plugins, MCP servers, and workflows that give engineers an AI-powered development environment spanning the full lifecycle — from brainstorming ideas to shipping pull requests.

Instead of configuring each tool individually, the Ontopix setup provides a single installation path that brings together 10 plugins, 9 MCP servers, and dozens of skills, commands, and agents into a cohesive workflow.

The Layered Architecture

The setup follows a three-layer model where each layer builds on the one below:

Layer 3: ontopix plugin       → Live Ontopix patterns via MCP
Layer 2: sanctum/pensive/...  → Generic dev workflows (commits, reviews, TDD)
Layer 1: Claude Code          → Core runtime

Layer 1 is Claude Code itself — the AI coding assistant that runs in your terminal.

Layer 2 adds generic development workflows through community and third-party plugins. These handle things like git commits, code review, debugging, and spec-driven development. They work in any project, regardless of your team's conventions.

Layer 3 is the Ontopix plugin. It connects Claude Code to Ontopix's documentation server (docs.ontopix.dev) via MCP, giving the AI live access to your team's engineering patterns, conventions, and standards. This layer makes Claude Code aware of how Ontopix does things.

The "Thin Plugin" Philosophy

The Ontopix plugin does not embed patterns or conventions in its source code. Instead, it configures an MCP connection to docs.ontopix.dev and provides two thin skills (/setup-repo and /check-compliance) that instruct Claude how to query patterns on demand.

This means:

• Conventions are always current. When a pattern is updated on docs.ontopix.dev, every Claude Code session sees the new version immediately.
• No staleness. There is no cached copy to go out of date.
• Minimal token footprint. Patterns are fetched only when needed, not injected into every conversation.

What You Get

After following the recommended setup, you get:

These tools cover the full development lifecycle:

• Brainstorming — Explore ideas collaboratively with structured design sessions
• Specification — Write PRDs and specs with guided workflows
• Architecture — Review system design and ADR compliance
• Planning — Break features into bite-sized implementation plans
• Implementation — Execute plans with TDD, subagent-driven development
• Code Review — Multi-dimensional review covering bugs, architecture, and tests
• Git and PR Workflows — Conventional commits, PR preparation, issue-driven development
• Debugging — Systematic root-cause analysis
• Resource Optimization — Token conservation and context management
• Live Pattern Access — Query Ontopix engineering standards in real time

Next Steps

• Getting Started — Install everything in minutes
• Available Tools — See every command, agent, and MCP server
• Use Cases — Learn workflows for real scenarios
• Plugins — Deep-dive into each plugin
• Contributing — Help improve the setup