init-deep

Hierarchical AGENTS.md documentation is generated by analyzing repository structure and applying scoring-based directory selection to create depth-aware files. This ensures AI agents can quickly locate necessary knowledge across nested codebases without manual mapping.

The best way to map repository complexity is using filesystem discovery combined with scoring heuristics like file counts and depth rules. This approach identifies exactly where documentation is needed, guiding AI agents through nested codebases efficiently.

Yes, hierarchical AGENTS.md generation supports incremental updates with built-in deduplication rules. These rules prevent child documents from repeating parent content, ensuring your project documentation remains navigable and free of redundant information during updates.

Yes, you can automate codebase discovery and project documentation tasks by triggering generation flows with user prompts. The system analyzes your repository structure and safely creates or updates AGENTS.md files based on detected complexity.

To reduce noise and redundant content when updating project documentation, use a full regeneration flow that applies deduplication rules. This removes general content from child documents, keeping your AGENTS.md files understandable and navigable for AI agents.