meta-loop-methodology

The meta-loop methodology is a repeatable AI-assisted workflow that uses external documentation artifacts like wireframes and specifications to achieve single-shot software delivery. It combines a Wireframe, Spec, and Implementation loop with Plan, Ask, and Agent modes to guide AI actions.

To generate code with an AI agent, you draw a wireframe, verify understanding, generate a specification, and then implement it. You capture learnings in documentation files to ensure the process is repeatable across sessions.

Yes, this methodology is designed for cross-platform prototyping and transferability. It emphasizes external documentation as the durable artifact, ensuring persistent learning and reproducible development across different platforms and projects.

Discarding and regenerating faulty artifacts enforces a discipline that rebuilds code from clarified source documents. This approach prevents carrying over errors and maintains the integrity of the single-shot delivery workflow.

Documentation-centric learning preserves context by relying on external artifacts like wireframes, specs, and AGENTS.md files. These documents act as the durable memory for the AI, capturing learnings that persist across sessions and platforms.

Plan, Ask, and Agent modes are operational states that guide AI actions within the development loop. They structure how the AI interacts with the Wireframe, Spec, and Implementation phases to ensure controlled and repeatable software generation.