Task Decomposition

Decompose around execution and evidence boundaries, not arbitrary checklist length.

Procedure

Define the observable outcome and the end-to-end evidence that proves it.
Record the target, required inclusions, negative constraints, dependencies, approvals, and irreversible steps.
Split the work into independently verifiable units.
Order units by dependency and feedback speed.
Choose the smallest reversible first action that produces new information.
Expose only Now, Next, and Later unless the user requests the full plan.
Re-plan from evidence when a unit fails. Do not repeat the same attempt unchanged.
Reconcile every unit and constraint before a terminal claim. A completed unit, build, or lower-layer test proves only itself.
Before a long or effectful unit, externalize a restart packet: outcome, last verified state, current action, blocker, remaining authority, and exact re-entry step.
On resume, refresh live state and authority before acting. Do not assume volatile context survived or reuse a one-shot approval.

Check the task for: