What does this skill do, and when should you use it?
This is a prompt-engineering skill for UI/UX prototyping. It gathers requirements through incremental questioning — app type, platform, design preferences, technical constraints — selects one of four design systems (WeChat Work, iOS Native, Material Design, Ant Design Mobile), then builds a comprehensive prompt from a nine-section template covering hex color palettes, component dimensions, layout trees, Tailwind config, and realistic sample content. The output is prompt text, not prototype code; it must be passed to another AI tool or developer. The workflow includes a multi-category quality checklist and iterative refinement.
- Collects requirements through incremental questions (2-3 at a time) covering app type, target platform, viewport sizes, design style, framework preferences, and real content needs
- Loads the chosen design system's palette, component specs, and interaction patterns from bundled references/design-systems.md
- Assembles the prompt using the nine-section template in references/prompt-structure.md (role, task, tech stack, visual specs, implementation details, Tailwind config, content hierarchy tree, special requirements, output format)
- Replaces all placeholders with concrete values — hex codes like #3478F6, exact dimensions like 44px touch targets and 64px list rows, and realistic business content
- Self-checks against completeness, clarity, specificity, realism, and technical accuracy checklists
- Presents the result, offers refinement options, and can export to Markdown or text files
- A product manager needs a spec-grade prompt an AI can execute to prototype a Chinese enterprise mobile work app (WeChat Work style)
- An indie developer wants a complete HTML/Tailwind prototype prompt for an iOS-style consumer app
- An outsourced team needs consistently formatted design documentation with hex codes, component dimensions, and page structure trees for multi-party collaboration
- A designer wants to validate whether an Ant Design Mobile form-heavy app layout works at a 375px viewport
- A founder describing a dashboard prototype to an AI tool wants to avoid vague instructions like 'use some blue'
- Users expecting runnable prototype code — this skill generates prompts only; a second execution step by another AI tool or developer is required
- Projects needing desktop web design systems (e.g., Ant Design desktop) — coverage is limited to WeChat Work, iOS Native, Material Design, and Ant Design Mobile
- Users who want one-shot results — the incremental questioning flow can take multiple rounds
How do you install this skill?
- This skill only generates prompt text, not prototypes; outputs must be handed to another tool or developer.
- Publisher is unverified, and there is no skill-level versioning or changelog; track updates via whole-repo releases.
- Referenced design-system specs (e.g., WeChat Work colors) may become outdated; verify against current official guidelines.
- Static review did not execute the skill; actual generation quality is independently unconfirmed.
- Local filesystem
references/design-systems.mdreferences/prompt-structure.md (bundled reference files)
The skill lives in the stellarlinkco/myclaude repository (a bundle of 11 skills).
npx (interactive install of the collection)
npx github:stellarlinkco/myclaudenpx (list individually installable skills)
npx github:stellarlinkco/myclaude --listAlternatively, download manually from GitHub and place the skills/prototype-prompt-generator folder into your Agent Skills client's directory (default ~/.claude/skills/ for Claude Code). The exact standalone-install command for this one skill is not documented in the source.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Create a prototype prompt for a WeChat Work style enterprise mobile work app
- Generate a detailed UI prototype prompt for an iOS-style e-commerce app at a 375px viewport using Tailwind CSS
- Help me write a structured prototype prompt for an Ant Design Mobile form-heavy app with realistic content
- I want to describe a dashboard prototype to an AI tool — turn vague notes like 'use some blue' into a full spec with hex codes and dimensions
In an Agent Skills-capable client such as Claude Code, trigger it with prompts like 'create a prototype prompt', 'generate UI specifications', or 'design a mobile app'. The skill asks grouped questions, recommends a design system, then outputs the full prompt. Note: the output is prompt text — pass it to another AI tool or developer to produce the actual HTML/CSS/React code.
What are this skill's strengths and limitations?
- Exceptionally specific output: every color has a hex code, every measurement has units, real content instead of Lorem Ipsum placeholders
- Bundled full specifications for four design systems, with recommendation logic based on user geography, platform, and app type
- Nine-section template plus five quality checklists produce consistent, reusable results
- Pure instruction-based skill — no scripts, network, or specific runtime required; highly portable
- Generates prompts only, not prototype code — requires a second execution step
- Design system coverage limited to WeChat Work, iOS Native, Material Design, and Ant Design Mobile; desktop web systems (e.g., Ant Design desktop) are not mentioned
- No test suite or real-world output validation in the source; final quality depends on the AI tool that executes the prompt
- The questioning flow can take multiple rounds; users wanting one-shot results may find it tedious
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Prototype Prompt Generator this page | 59 · Recommended | ★ 2.8k | 5mo ago | AGPL-3.0 |
| UI/UX Consultation Guide | 56 · Use with care | ★ 113 | 1mo ago | Apache-2.0 |
| Figma Design System Screen Builder ✓ OpenAI · Official | 46 · Use with care | ★ 28k | 4mo ago | — |
| Frontend Design Skill | 52 · Use with care | ★ 5.5k | 5mo ago | MIT |
| Web Design Engineer | 50 · Use with care | ★ 13k | 14d ago | MIT |
How did FollowSkills review this skill?
Skill is text-only prompt generation: no code execution, no sensitive permissions, no external effects; least privilege trivially satisfied. Deducted for lack of explicit confirmation/rollback statements and publisher attribution only at repo level (unverified publisher).
Instructions are self-consistent; SKILL.md and both reference files cross-reference correctly with clear workflow and trigger conditions. Static cap 10; deducted for no test or edge-case execution evidence specific to this skill, failure feedback relies on model behavior.
Triggers are specific, non-fit boundary explicitly declared (generates prompts, not prototypes), covers China-relevant design systems (WeChat Work, Ant Design Mobile), no network dependency so mainland-China reachability is fine. Deducted for no usage evidence backing trigger precision; spec currency undisclosed.
Well-layered docs (SKILL.md + references), troubleshooting and quality checklists included, repo-level AGPL-3.0 license clear. Deducted for no skill-level version, changelog, or maintenance responsibility; update path depends on whole-repo releases.
Claims to produce structured, directly usable prototype prompts with complete templates, but static review cannot verify actual output quality; cap 7. Deducted for lack of representative verified outputs or comparative-benefit evidence over manual prompt writing.
Claims are auditable against source files (templates and specs present), but no third-party execution evidence; CI tests cover only codeagent-wrapper and harness, unrelated to this skill. Cap 5; deducted for no reproducible output validation or cross-source corroboration.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →