Educational companion to the Context_Map agent-first repository template.
This guide explains the ideas, patterns, and principles behind Context_Map. It is not operational documentation -- it does not tell you what to do in this specific project. Instead, it teaches you why the template is structured the way it is and how to apply the same thinking to your own repositories.
Think of it as a textbook that ships alongside the working example. The repo itself demonstrates every pattern; the guide explains the reasoning.
- Developers adopting agent-led development for the first time
- Teams transitioning from human-centric repos to agent-first structures
- AI agents that need to understand the methodology (though agents should
prefer
docs/for operational questions) - Architects evaluating whether Context_Map fits their workflow
- Sequential for learning. Chapters build on each other. Start at Chapter 1 and read through to understand the full methodology.
- Individual chapters for reference. Each chapter is self-contained enough to be useful on its own. Jump to whichever topic you need.
Every chapter follows the same structure: Problem (what goes wrong without this pattern), Principle (the core idea), Implementation (how to do it), Adaptation (how to tailor it to your context).
| Chapter | Title | Description |
|---|---|---|
| 01 | Why Agent Legibility Matters | The case for structuring repos so AI agents can read them effectively |
| 02 | Progressive Disclosure | Layering information so agents find what they need in three hops or fewer |
| 03 | Multi-Agent Setup | Supporting Claude Code, Cursor, Copilot, and other agents from one repo |
| 04 | Execution Plans as First-Class Artifacts | Giving complex tasks structured, lifecycle-managed documents that agents can discover and follow |
| 05 | Quality Scoring and Enforcement | Defining quality as measurable dimensions with automated checks and debt tracking |
| 06 | Documentation Gardening | Treating docs as a living system with freshness detection and continuous maintenance |
| 07 | Session Handoffs for Continuity | Capturing session state so the next agent resumes without loss of context |
| 08 | Building Claude Code Skills | Encoding repeatable workflows as slash-command skills with instructions and reference data |
The guide/ directory is reference material. Agents must never modify these
files during normal operation. If a guide chapter needs updating, that is a
deliberate editorial decision -- not something that happens as a side effect of
working on the project.
Operational truth lives in docs/. Educational content lives here.
This guide is part of the Context_Map agent-first repository template.