Project level DB inside the project

[brainless]

We want to change the structure of our Project level data, moving them away from the central SQLite DB. Instead we want to move all Project level data into a .nocodo folder inside the project itself. The folder will have NDJSON files to organize the data, starting with a project.ndjson.

Here is the current state of the entities that are Project related and are in SQLite. We mark the items which should stay in SQLite after this migration. Rest of the items should all use the .nocodo folder in the project.

Project Basic Information

• Unique ID - stays in SQLite
• File system path - stays in SQLite
• Name
• Primary programming language (optional)
• Framework (optional)
• Current status
• Creation timestamp
• Last update timestamp
• Technologies (JSON-serialized list of detected technologies)

Project Components (sub-apps within project)

• Component ID
• Component name
• Relative path within project
• Programming language
• Framework (optional)
• Creation timestamp

Works (development sessions/conversations)

• Work ID
• Title
• Associated tool name (optional)
• Status
• Creation timestamp
• Update timestamp

Related AI Activity (indirectly through works)

• AI sessions with tool usage tracking
• Work messages (user/AI conversation history)
• AI session outputs
• LLM agent sessions and tool calls

The core project tracking centers on metadata, technology detection, and development activity history through the works system.