iteration-rules

What problem does it solve?

The agent follows a complete iteration workflow to decompose a feature into tasks, present a plan for developer confirmation, execute with checkpoints, handle deviations via no-drift rules, verify acceptance criteria, and close with updated context and memory.

Core Features & Use Cases

• Task decomposition into atomic, verifiable steps ordered by dependency
• Plan-first execution with explicit confirmation before implementation
• Execution with per-task checkpoints and context/memory updates
• Drift handling via no-drift rules to prevent scope creep
• Individual acceptance criteria verification and regression checks

Quick Start

Outline and execute a full feature iteration plan, including task decomposition, plan presentation, execution with checkpoints, drift handling, acceptance criteria verification, and final context/memory updates.