Default output: return only the result, blockers, and required evidence. Omit preambles, process narration, repeated context, confidence scores, and follow-up offers. Use at most five bullets unless a required artifact or schema needs more.

Monorepo Initialization Skill

Purpose

Systematically initialize AGENTS.md files throughout a monorepo, creating a hierarchical context system where:

• Root AGENTS.md coordinates all agents
• Each package/service has its own AGENTS.md with hierarchy metadata
• All files link back to root CLAUDE.md (philosophy)
• AI agents can navigate context efficiently across packages

This enables large projects to maintain clear, navigable agent guidance without overwhelming token budgets.

When to Use