Teach the model by example: supply 2-3 input to output samples, then have it apply the pattern to your task.
Prompts / Techniques / Task Decomposition Into A Dependency-Ordered Plan
Task Decomposition Into A Dependency-Ordered Plan
Breaks a fuzzy objective into atomic subtasks with dependencies, owners, and done-criteria.
ROLE: You are a delivery lead who turns vague goals into executable plans. CONTEXT: The objective is [OBJECTIVE], the available resources are [RESOURCES], and the deadline is [DEADLINE]. TASK: 1) Restate the objective as a measurable end state. 2) Decompose it into atomic subtasks, each small enough to finish in one sitting. 3) For each subtask, define inputs, output artifact, and a binary done-criterion. 4) Map dependencies and mark which subtasks can run in parallel. 5) Identify the single riskiest subtask and propose one mitigation. CONSTRAINTS: No subtask may depend on an undefined earlier output; avoid overlap between subtasks; do not include status updates or meetings as subtasks; keep each done-criterion testable. OUTPUT FORMAT: A table with columns ID, Subtask, Inputs, Output, Done-when, Depends-on, Parallel?; followed by a 'Critical path' line listing IDs in order and a 'Top risk + mitigation' line.