clarification

Clarification

Purpose

Gather everything needed before any planning begins. One organized batch of questions — not a back-and-forth interrogation.

Question Categories (cover all that apply)

1. Scope: What is in scope? What is explicitly out of scope?
2. Success criteria: What does done look like? How will we know it works?
3. Tech stack: Language, framework, runtime constraints. Existing conventions to follow?
4. CI/CD assumption: Assume in place unless user says otherwise. Confirm only if the request touches deployment.
5. Manual validation: Does this touch a UI? Major API surface changes? External integration impact? These will require manual review.
6. Existing codebase: Is this greenfield or modifying existing code? If existing, where does it live?
7. Constraints: Timeline, performance, security, compliance requirements?

Output: Clarification Summary

After getting answers, produce a structured summary:

## Clarification Summary

**Request:** <one-sentence restatement>
**Scope:** <what's in, what's out>
**Success criteria:** <observable outcomes>
**Tech stack:** <languages, frameworks, versions>
**Manual validation required:** <yes/no, what triggers it>
**Assumptions confirmed:** <list of things user explicitly confirmed>
**Out of scope confirmed:** <list of things user explicitly excluded>

Ask the user: "Does this summary accurately capture your request? Confirm to proceed to planning."

Do not proceed until user explicitly confirms. Write the confirmed summary to memory under key clarification_summary.

Rules

• Ask all questions in one message, organized by category
• Do not ask about things already stated in the initial request
• Every item in the summary must have been stated or confirmed by the user — no inferences