deepinit

To initialize an agent harness, the deepinit skill scans your codebase and automatically generates essential documentation like AGENTS.md, ARCHITECTURE.md, and a docs/ directory structure to ensure disciplined AI-assisted engineering.

Generating architecture documentation is handled by analyzing the codebase to identify key components, which then automatically outputs a structured ARCHITECTURE.md file and a comprehensive docs/ knowledge base.

Setting up project boundaries and coding standards for AI agents is achieved by creating an AGENTS.md file during project initialization, which establishes the rules and structure required for effective agent collaboration.

You do not need to manually create a docs directory, because the project initialization process automatically scans the codebase and creates the directory structure alongside the essential documentation files.

You can use codebase analysis to establish a standardized project structure, as the process scans your existing code to generate the necessary documentation and directory structures for effective AI-driven development.