planner

To break down complex software development objectives, you can generate a structured markdown plan that decomposes large goals into actionable tasks. This process includes defining clear file modifications, verification steps, and done criteria to ensure autonomous execution by AI agents.

Goal-backward verification is a planning mechanism that ensures task breakdowns meet defined success criteria by working backward from the desired outcome. It structures dependencies and verification steps so that the final generated plan aligns directly with the original objective.

Structuring task dependencies for execution waves requires analyzing and mapping relationships between individual actionable tasks within your plan. By identifying these dependencies, you can organize tasks into sequential or parallel waves that maintain logical flow and maximize execution efficiency.

Yes, you can generate a PLAN.md file that outlines exact steps, required file modifications, and verification methods for each task. This structured markdown format provides a clear path to completion for new features by detailing specific actions and done criteria.

Planning for AI agent workflows supports context budget management by defining discovery levels within the generated plan. Managing the context budget ensures the AI agent maintains high quality and performance during autonomous execution without exceeding processing limits.