writing-plans

To automate plan generation, the writing-plans skill reads current-task/definition.md and doctrine files like constitution.md to produce a formal current-task/plan.md contract. This output preserves MVP scope and records risks, acceptance criteria, and task decomposition for downstream verification.

Maintaining task plan continuity requires generating a durable plan.md contract that survives model switches and session resets. By converting task definitions and operational doctrine into a formal plan file, downstream execution components can reliably resume work without losing scope or context.

No, the plan generation workflow explicitly does not create plan_context.md. It strictly produces current-task/plan.md as the formal contract, delegating contextualization to a separate plan-contextualize step to keep planning and execution responsibilities distinct.

You can enforce operational limits and constitution rules by feeding limits.md and constitution.md into the plan generation process. The skill reads these doctrine files alongside the task definition to ensure the generated plan.md respects guardrails and preserves MVP scope.

You need a formal plan contract when handing off tasks to downstream execution components or cross-session workflows. The generated plan.md records task decomposition, acceptance criteria, and risks, providing a durable agreement that guides and verifies execution.

Yes, plan generation supports workflows driven by powerful models like Claude Code plan mode. It generates a plan that survives model switches and session resets, ensuring the formal plan.md contract remains intact across different execution environments.