agent-onboarding

To set up a repository for autonomous agent operation, run an automated 7-phase onboarding process that creates essential context files like CLAUDE.md, INTENT.md, and SPECS/. This systematically defines the project's goals, context, and recurring tasks for the agent environment.

A self-improving agent runtime is a continuous loop installed during repository onboarding that monitors for friction, logs errors, and generates improvement proposals. It evolves the environment over time by automatically detecting and proposing solutions for context gaps within the repository.

Yes, you can onboard an existing codebase for AI agent context engineering. The onboarding process supports both greenfield projects and existing codebases, enforcing context best practices to ensure autonomous agents have the necessary specifications and intent.

To configure context files like CLAUDE.md and INTENT.md, run the automated onboarding workflow which guides users through a 7-phase process. This creates the essential files and specifications required to define the agent's goals and operational context.

Yes, autonomous agent onboarding supports optional git worktree support for behavior profiles. This allows you to isolate different agent behaviors or context configurations within separate worktrees while maintaining the self-improving runtime environment.