What does this skill do, and when should you use it?
This skill rewrites rough UI concepts into specific, structured prompts for Google Stitch. It fills in missing platform, page type, layout structure, visual direction, color roles, and UI component terminology. When a project contains DESIGN.md, it reads the documented palette, typography, and component styles and adds them as a required design-system section. Without that file, it appends guidance to create one for consistent multi-screen designs.
Reads the user's UI concept and assesses missing information; requires consulting the official Stitch prompting guide before enhancement; checks the current project for DESIGN.md and reads its design-system content when present; replaces vague language with concrete UI/UX terms, page sections, visual descriptors, and functional color definitions; returns a copyable enhanced prompt by default, or writes next-prompt.md or a user-specified filename when requested.
- A product designer can use it before sending a vague request such as “make a settings page” to Stitch.
- A frontend developer can use it to restructure a prompt after Stitch produces an unsatisfactory first result.
- A team maintaining multiple screens can use it to inject the shared DESIGN.md context into new generation prompts.
- A user making a narrowly scoped change can use it to create a targeted-edit prompt that preserves existing elements.
How do you install this skill?
- The skill requires consulting external Stitch documentation but provides no fallback when that documentation or service is unreachable; results may be less stable in restricted environments.
- Write permission and writable-file scope are not narrowly specified; explicit user confirmation and a recoverable copy should precede file creation.
- Examples are primarily English and do not establish support for Chinese UI prompts, Chinese design terminology, or mainland-China network conditions.
- Network access
- Local filesystem
- MCP Server
Stitch MCP server
This skill is part of the stitch-utilities plugin in google-labs-code/stitch-skills. The README documents selective installation with npx skills add google-labs-code/stitch-skills; it does not document a dedicated command for installing only enhance-prompt. The Stitch MCP server must also be configured and running in the agent environment.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/google-labs-code/stitch-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/plugins/stitch-utilities/skills/enhance-prompt" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Enhance this prompt: make a settings page
- Add a search bar.
After installing the skill and configuring Stitch MCP, submit a request such as “Enhance this prompt: make a settings page” or “Add a search bar.” The skill returns a structured enhanced prompt. To save it, explicitly request next-prompt.md or provide a custom filename.
What are this skill's strengths and limitations?
- Provides a defined four-stage workflow covering assessment, design-system checking, enhancement, and formatting.
- Translates vague descriptions into concrete terms such as navigation bars, primary CTAs, card grids, and labeled input fields.
- Uses DESIGN.md when available to preserve palette, typography, and component-style context across screens.
- Supports both complete-page prompts and focused targeted-edit prompts.
- Requires a configured and running Stitch MCP server, which is not included in the repository.
- Requires consulting an external Stitch prompting guide, so network availability affects the workflow.
- The SKILL.md provides no test suite, version compatibility matrix, or failure-handling specification.
- When DESIGN.md is absent, the skill only recommends creating one; it does not extract a design system from an existing project.
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 |
|---|---|---|---|---|
| Stitch Prompt Refiner this page | 49 · Use with care | ★ 8.5k | 2mo ago | Apache-2.0 |
| Stitch Design Taste | 42 · Not recommended | ★ 8.5k | 2mo ago | Apache-2.0 |
| Stitch Screen Generator | 45 · Not recommended | ★ 8.5k | 2mo ago | Apache-2.0 |
| Stitch Design System Manager | 45 · Not recommended | ★ 8.5k | 2mo ago | Apache-2.0 |
| Stitch Design System Synthesizer | 45 · Not recommended | ★ 8.5k | 2mo ago | Apache-2.0 |
How did FollowSkills review this skill?
The skill declares only Read and Write tools and defaults to returning text, with file output described as optional and user-requested; no malware, credential theft, or covert exfiltration is evident. However, the permissions are broad and there is no explicit write confirmation, rollback plan, data-flow disclosure, or boundary for external documentation access, so points are deducted.
The workflow and output template are broadly self-consistent, covering input assessment, DESIGN.md lookup, enhancement, and optional file output. There are no tests, scripts, abnormal-input handling, failure feedback, or diagnostic behavior for failed inference, so static evidence supports only a modest score.
Trigger scenarios, input types, and main enhancement steps are reasonably clear for UI prompt polishing. Non-fit boundaries, language support, and precise exclusion conditions are not defined, and the Stitch ecosystem plus official documentation are dependencies whose reachability from mainland-China networks is not established, so points are deducted.
SKILL.md, README.md, and the keyword reference provide a readable progressive structure with installation notes, examples, and output conventions. The skill lacks its own versioning, changelog, maintenance owner, update path, FAQs, and explicit limitations; license and governance evidence mainly comes from shared repository files, so points are deducted.
The skill clearly produces a copyable, structured Stitch prompt using design-system guidance, UI vocabulary, and page organization. The examples support completion of the core task, but there is no real generated-result or comparative evidence, and user review may still be needed; the static ceiling therefore limits the score.
The source exposes its rules, templates, examples, and dependency declaration, providing limited auditability. There is no test suite, CI coverage for this skill's key paths, or third-party execution evidence, so independent verification of output quality is limited.
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 →