Deep Init Skill

Creates comprehensive, hierarchical AGENTS.md documentation across the entire codebase.

Core Concept

AGENTS.md files serve as AI-readable documentation that helps agents understand:

• What each directory contains
• How components relate to each other
• Special instructions for working in that area
• Dependencies and relationships

Hierarchical Tagging System

Every AGENTS.md (except root) includes a parent reference line:

**Parent context:** `../AGENTS.md`

This MUST be visible prose, never an HTML comment. Claude Code strips HTML comments before a memory file reaches the model, so <!-- Parent: ... -->, <!-- Generated: ... --> and a commented MANUAL boundary are invisible to the agent that is supposed to act on them.

This creates a navigable hierarchy:

/AGENTS.md                          ← Root (no parent line)
├── src/AGENTS.md                   ← **Parent context:** `../AGENTS.md`
│   ├── src/components/AGENTS.md    ← **Parent context:** `../AGENTS.md`
│   └── src/utils/AGENTS.md         ← **Parent context:** `../AGENTS.md`
└── docs/AGENTS.md                  ← **Parent context:** `../AGENTS.md`

Loading Model

Stock Claude Code discovers nested CLAUDE.md by basename and never loads a nested AGENTS.md; a root @AGENTS.md import does not reach subdirectories either. The nested files generated here are delivered by OMC's PostToolUse directory-context injector, which walks up from each accessed file and injects the nearest AGENTS.md/README.md once per session.

That means the hierarchy requires OMC's hooks to be active. When OMC hooks are disabled (DISABLE_OMC=1, OMC_SKIP_HOOKS=post-tool-use) or the plugin is not registered, add a root-level CLAUDE.md symlink or @AGENTS.md import for the root file and expect nested files to stay unread.

AGENTS.md Template

# {Directory Name}