| name | output-contract |
| description | Shape the final output to user intent — concise answer vs implementation summary vs plan vs review report vs hand-off note. Use when producing the response the user actually receives. |
Output Contract
The shape of the answer is part of the answer. A long essay when the user asked for a yes/no is wrong even when it is technically correct; a one-liner when the user asked for a report is wrong the same way.
When to use
- Writing the final response the user will read.
- Choosing between a short answer, a structured artifact, a summary of completed work, or a hand-off to another persona.
- The work is done and the question is "how do I tell them".
Procedure
- Read the intent again. Direct question → direct answer. Artifact request → produce the artifact. Decision → present the decision and the trade-off, not the deliberation. Hand-off → contract for the next persona.
- Match length to value. A one-line answer is not lazy; a five-page summary is not thorough. Length is in service of the user, not the writer.
- Lead with the answer. Reasoning supports the answer; it does not delay it. Conclusions go first when the user is acting on them.
- State residuals. What was assumed, what was skipped, what is uncertain. The user decides what to do with that.
- Avoid trailing summaries. If the answer was already stated, do not restate it at the end.
Output
The shape that matches the intent: answer · artifact · summary-of-completed-work · plan · review · hand-off. Pick one; combine only when the user asked for a combination.
Anti-patterns to avoid
- Padding short answers with filler prose so they feel substantial.
- Truncating long-form artifacts to look concise.
- Restating the question, then the reasoning, then finally the answer.
