init-deep

To generate AGENTS.md context files, you run the init-deep skill against your project. It uses concurrent exploration agents, bash analysis, and LSP codemaps to automatically create hierarchical documentation starting from the root directory down to complexity-scored subdirectories.

Providing AI context for an existing codebase is best achieved by generating structured AGENTS.md files. This approach uses a scoring matrix to identify and document critical areas, preventing information overload and ensuring the AI agent operates with precise, context-aware information.

Complexity scoring for project documentation initialization evaluates subdirectories using concurrent exploration agents and LSP codemaps. This scoring matrix identifies critical areas within the codebase, ensuring that hierarchical AGENTS.md files are only generated where documentation is genuinely needed.

You can update existing AGENTS.md files without a full regeneration. The initialization process supports both initial setup and targeted updates, offering options for full regeneration and depth control to maintain your hierarchical codebase documentation efficiently.

You do not need any external dependencies to run project context generation. The skill operates independently using built-in bash structural analysis, LSP codemaps, and concurrent exploration agents to analyze your codebase and produce the required AGENTS.md documentation.

You should not use automated codebase analysis for AI context if your project lacks sufficient structural complexity or code volume. Because the documentation generation relies on a complexity scoring matrix, extremely flat or minimal projects may not benefit from the hierarchical AGENTS.md output.