agents-md-authoring-majo

What problem does it solve?

AGENTS.md files are often inconsistent, bloated, or omitted, causing agents to operate with incomplete or conflicting context. This guide provides a structured, concise approach to authoring AGENTS.md that consistently delivers high-signal context for AI coding agents and reduces context-switching.

Core Features & Use Cases

• Six-Core-Areas Framework: Commands, Testing, Project Structure, Code Style, Git Workflow, Boundaries, with guidance on exact commands, testing expectations, high-signal structure, concrete examples, and safe boundaries.
• Progressive Disclosure: Keeps root guidance compact while linking to deeper docs and project-specific details.
• Three-Tier Boundaries: ALWAYS/ASK FIRST/NEVER to reduce ambiguity and enforce safety.
• Monorepo support: Nested AGENTS.md files enable project-specific guidance while preserving global rules.
• Iterative Refinement: Track agent mistakes and update guidance to steadily improve performance.
• Cross-Skill Integration: Works with skill-authoring-majo, task-planning-majo, and git-majo for end-to-end governance.

Quick Start

Create a concise AGENTS.md for a project that uses the six-core-areas framework and three-tier boundaries, citing real code examples from the repo and updating after agent mistakes.