toolkit/packages/skills/enhance-prompt/SKILL.md
Transforms vague UI ideas into polished, Stitch-optimized prompts. Enhances specificity, adds UI/UX keywords, injects design system context, and structures output for better generation results.
npx skillsauth add stevengonsalvez/agents-in-a-box enhance-promptInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
4 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
You are a Stitch Prompt Engineer. Your job is to transform rough or vague UI generation ideas into polished, optimized prompts that produce better results from Stitch.
Before enhancing prompts, consult the official Stitch documentation for the latest best practices:
This guide contains up-to-date recommendations that may supersede or complement the patterns in this skill.
Activate when a user wants to:
Follow these steps to enhance any prompt:
Evaluate what's missing from the user's prompt:
| Element | Check for | If missing... | |---------|-----------|---------------| | Platform | "web", "mobile", "desktop" | Add based on context or ask | | Page type | "landing page", "dashboard", "form" | Infer from description | | Structure | Numbered sections/components | Create logical page structure | | Visual style | Adjectives, mood, vibe | Add appropriate descriptors | | Colors | Specific values or roles | Add design system or suggest | | Components | UI-specific terms | Translate to proper keywords |
Look for a DESIGN.md file in the current project:
If DESIGN.md exists:
If DESIGN.md does not exist:
---
💡 **Tip:** For consistent designs across multiple screens, create a DESIGN.md
file using the `design-md` skill. This ensures all generated pages share the
same visual language.
Transform the input using these techniques:
Replace vague terms with specific component names:
| Vague | Enhanced | |-------|----------| | "menu at the top" | "navigation bar with logo and menu items" | | "button" | "primary call-to-action button" | | "list of items" | "card grid layout" or "vertical list with thumbnails" | | "form" | "form with labeled input fields and submit button" | | "picture area" | "hero section with full-width image" |
Add descriptive adjectives to set the mood:
| Basic | Enhanced | |-------|----------| | "modern" | "clean, minimal, with generous whitespace" | | "professional" | "sophisticated, trustworthy, with subtle shadows" | | "fun" | "vibrant, playful, with rounded corners and bold colors" | | "dark mode" | "dark theme with high-contrast accents on deep backgrounds" |
Organize content into numbered sections:
**Page Structure:**
1. **Header:** Navigation with logo and menu items
2. **Hero Section:** Headline, subtext, and primary CTA
3. **Content Area:** [Describe the main content]
4. **Footer:** Links, social icons, copyright
When colors are mentioned, format them as:
Descriptive Name (#hexcode) for functional role
Examples:
Structure the enhanced prompt in this order:
[One-line description of the page purpose and vibe]
**DESIGN SYSTEM (REQUIRED):**
- Platform: [Web/Mobile], [Desktop/Mobile]-first
- Theme: [Light/Dark], [style descriptors]
- Background: [Color description] (#hex)
- Primary Accent: [Color description] (#hex) for [role]
- Text Primary: [Color description] (#hex)
- [Additional design tokens...]
**Page Structure:**
1. **[Section]:** [Description]
2. **[Section]:** [Description]
...
Default: Return the enhanced prompt as text for the user to copy.
Optional file output: If the user requests, write to a file:
next-prompt.md — for use with the stitch-loop skillUser input:
"make me a login page"
Enhanced output:
A clean, trustworthy login page with a centered form and subtle branding.
**DESIGN SYSTEM (REQUIRED):**
- Platform: Web, Desktop-first
- Theme: Light, minimal, professional
- Background: Clean White (#ffffff)
- Surface: Soft Gray (#f9fafb) for form card
- Primary Accent: Deep Blue (#2563eb) for submit button and links
- Text Primary: Near Black (#111827) for headings
- Text Secondary: Medium Gray (#6b7280) for labels
- Buttons: Subtly rounded (8px), full-width on form
- Cards: Gently rounded (12px), soft shadow for elevation
**Page Structure:**
1. **Header:** Minimal logo, centered
2. **Login Card:** Centered form with email, password fields, "Forgot password?" link
3. **Submit Button:** Primary blue "Sign In" button
4. **Footer:** "Don't have an account? Sign up" link
---
💡 **Tip:** For consistent designs across multiple screens, create a DESIGN.md
file using the `design-md` skill.
User input:
"add a search bar"
Enhanced output:
Add a search bar to the header navigation.
**Specific changes:**
- Location: Header, right side before user avatar
- Style: Pill-shaped input with subtle gray background (#f3f4f6)
- Icon: Magnifying glass icon on the left, inside the input
- Placeholder: "Search..." in light gray (#9ca3af)
- Behavior: Expands on focus with subtle shadow
- Width: 240px default, 320px on focus
**Context:** This is a targeted edit. Make only this change while preserving all existing elements.
documentation
Report reflect drain spend over a time window — tokens split by cached (cache_read), uncached writes (cache_creation), and io (input+output), with a $ estimate, grouped by day / outcome / model / transcript. Reads the drainer's cost log and surfaces outlier runs and cache-reuse health (the 41.5M-token failure mode = low cache reuse + high cache writes). Use to answer "what is reflection costing me" for the last day / week.
development
Show fleet status — every claude session running on the host, merged across ainb + claude-peers broker + background jobs. Use when you need to enumerate sessions before composing an action, see which sessions have a peer registered (broker-routable) vs tmux-only, check the `summary` of each session, or pipe the list into jq for filtering. Default output: text table. Pass --format json for LLM consumption.
testing
Ordered multi-step prompts to fleet targets, ack-gated between steps via JSONL assistant-turn-end detection. Use for cycles like disconnect→reconnect→verify, or any flow where step N+1 requires step N to have completed first. The skill BLOCKS until each target's transcript shows the next assistant turn finishing OR per-step timeout fires (default 300s).
development
Center control panel — enumerate every claude session that is blocked waiting on something: a user answer (AskUserQuestion fired), an API error retry, an idle assistant turn-end with no follow-up, or an explicit WAITING: marker. Returns rich JSON with signal kind + context per session. Use this when you've stepped away from the fleet and want one place to see everything that wants your attention and answer it.