Prompt Writer
Transform messy thoughts into clear, effective prompts
Claude CodeGeneric
---
description: Transform messy thoughts into clear, effective prompts
---
# Prompt Writer
Transform unstructured text into a well-organized prompt.
## Input
$ARGUMENTS
## Your Task
Analyze the above input and produce a clean, effective prompt.
### Step 1: Understand Intent
Identify from the user's text:
- **Goal**: What are they trying to accomplish?
- **Context**: What background is relevant?
- **Constraints**: Any limitations or requirements?
- **Scope**: How large is this change? Single file? Multi-file? Multi-session?
- **Verification**: How will Claude verify the work is complete and correct?
### Step 2: Clarity Check (Interview if Needed)
**If any of these are unclear, invoke the `interview` skill:**
| Signal | Action |
|--------|--------|
| Goal is vague ("improve X", "fix the thing") | Interview: clarify objective |
| Scope is ambiguous ("add auth", "refactor") | Interview: define boundaries |
| Multiple valid interpretations | Interview: resolve ambiguity |
| Verification can't be inferred | Interview: define success criteria |
| Design decisions implicit | Interview: surface choices |
**How to interview:**
- Use `AskUserQuestion` tool with 1-4 targeted questions
- Offer smart options (mark one as "Recommended" when appropriate)
- Users can always select "Other" for custom input
- Iterate until requirements are clear
**After interview, inject results as context:**
```markdown
## Interview Results
**Goal**: [Clarified objective]
**Scope**: [What's in/out]
**Design Decisions**: [Choices made]
**Verification**: [Success criteria]
Then continue to Step 3 with refined understanding.
Step 3: Identify Verification Steps
Critical: Clear verification steps are essential.
Consider:
- How will Claude know the task is complete?
- What tests should pass?
- What commands verify success?
- What should Claude check before declaring done?
If verification is unclear from the input, ASK the user:
"How should Claude verify this work is complete? For example: run tests, check specific behavior, validate output format, etc."
Step 4: Output
Output the refined prompt in this format:
## Goal
[Clear statement of objective]
## Context
[Relevant background, constraints, environment]
## Requirements
- [Requirement 1]
- [Requirement 2]
## Verification
- [How to verify it's done]
Guidelines
- Preserve the user's intent; don't add scope
- Be concise but complete
- Include constraints the user implied but didn't state explicitly
- Always include verification steps - ask if unclear
Maintain Prompt Writer?
Let people know it's listed here — add the badge (live metrics, light/dark aware) or a plain link to your README or docs.
[Prompt Writer on getagentictools](https://getagentictools.com/loops/thebrettbarlow-prompt-writer?ref=badge)