ralph-wiggum

Spec-driven development is automated by applying iterative loops with fresh context per iteration, ensuring verifiable completion across multiple specs. The agent tests and commits changes only when all acceptance criteria are fully satisfied.

State persistence is managed by storing specs, plans, and history directly on disk for reproducibility and resumption. This approach ensures that autonomous AI coding loops maintain auditable progress across multiple development iterations.

Set up a spec-driven loop by defining a set of specs with explicit acceptance criteria for your multi-spec software project, then instruct your AI to run the autonomous loop. It processes tasks incrementally and signals DONE only after all criteria pass.

Yes, autonomous AI coding supports incremental, verifiable completion across tasks for multi-spec software projects. It applies isolated per-iteration context and on-disk state persistence to ensure clean progress through each spec.

The autonomous AI agent signals DONE only after acceptance criteria pass to enforce guarded autonomy and prevent incomplete task execution. This explicit signaling ensures changes are pushed and verified before moving to the next iteration.

Yes, the spec-driven development loop requires fresh context per iteration to ensure clean progress across multiple specs. This isolation prevents context bleed and maintains reproducible, auditable state on disk for incremental task execution.