| name | plan-output |
| description | Shape implementation plans into a concise, execution-ready format with problem, scope, acceptance criteria, risks, steps, and verification. Use in Plan mode or when the user asks for a plan. |
Plan output
Goal
Turn planning work into an actionable artifact which is easy to review and execute.
Use this in plans for
- Features
- Refactors
- Migrations
- Debugging
Default shape
Use this structure unless the user asks for something else:
-
Overview
- One short paragraph describing changes and providing reasoning.
-
Scope
- In scope: affected areas, files, systems, or behaviors.
- Out of scope: anything intentionally excluded.
-
Acceptance criteria
- 2-5 testable bullets defining "done".
-
Risks
- Key failure modes, trade-offs, dependencies, or rollout concerns.
-
Implementation steps
- Ordered steps from first change to last.
- Prefer smallest shippable increments.
-
Verification
- Tests, checks, or manual validation steps.
Output rules
- Keep plans concise by default
- Prefer concrete nouns over vague phrases
- Make each step observable and actionable
- Call out assumptions when requirements are unclear, ask follow up questions to gain clarity
- Include "out of scope" when it prevents plan creep
- Do not pad with obvious filler
Example skeleton
## Problem
<what is changing and why>
## Scope
- In scope: ...
- Out of scope: ...
## Acceptance Criteria
- ...
- ...
## Risks
- ...
## Implementation Steps
1. ...
2. ...
## Verification
- ...