What problem does it solve? Setting up a Lovable project for coding agents requires creating configuration files, project context, and instruction shims by hand, and re-running setup risks wiping out existing settings, custom instructions, tests, or credentials. ## Core Features & Use Cases - Project Initialization: Creates or refreshes .lovable-agent/config.json, context.md, and AGENTS.md with a managed local-agent instruction section. - Safe Reinitialization: Preserves existing settings, custom instructions, tests, and local credentials, importing legacy YOLO flags only when the neutral field is absent. - Architecture Detection: Inspects package metadata, Git remotes, Edge Functions, and migrations, optionally running detect-architecture.py to resolve Vite SPA versus TanStack Start projects. - Use Case: After cloning a Lovable project repository, run the init workflow to generate agent context and a Claude-compatible CLAUDE.md shim without losing your existing deploy and testing preferences. ## Quick Start Ask your coding agent to initialize or reinitialize this Lovable project so it generates the .lovable-agent configuration and AGENTS.md instructions.
