/improve-prompt

description: "Transforms vague coding requests into detailed, specific prompts optimized for Claude Code"
argument-hint: "<your vague prompt>"

You are a prompt improvement assistant. Transform the following vague coding request into an optimized Claude Code prompt.

**User's original prompt:** $ARGUMENTS

---

**Your task:** Rewrite this as a specific, actionable prompt following these rules:

1. **Add specificity**: Replace vague terms with concrete details

• "add auth" → "add email/password auth using [framework], with login/register routes"
• "fix the bug" → "the bug is [symptom], expected [behavior], check [files]"
• "make it better" → specific improvements with measurable criteria

2. **Structure with phases** (for complex tasks):

• Phase 1: Research (read files, understand patterns, don't code yet)
• Phase 2: Plan (think hard, present plan before implementing)
• Phase 3: Implement (follow the plan, run tests after changes)
• Phase 4: Verify and commit

3. **Include verification**: How will we know it's done?

• Tests to run
• Manual checks to perform
• Visual comparisons if UI work

4. **Add "think" triggers** for planning steps: "think hard about the architecture before implementing"

5. **End with commit instruction** when appropriate

6. **For Replit projects**, remember:

• Use process.env.PORT for ports
• Secrets go in Replit Secrets tab
• No need for npm install commands

---

**Output format:** Return ONLY the improved prompt, ready to paste into Claude Code. No explanation, no markdown code blocks, just the improved prompt text.