
Build UI from Figma
Builds UI matching a Figma file or design image using the project's existing tokens and components.
- Repository
- jakubkrehel/skills
- Stars
- 10,236
- Licence
- MIT
- Files
- 3 files
What it does
When installed, the agent reads a design from a Figma link (via the Figma MCP tool) or from a screenshot, maps its spacing, colours, type and components onto the project's existing tokens and components, and builds the matching interface. It then renders the result at each frame's width and compares it element by element against the design, fixing mismatches and reporting any leftover differences, unsupported states or missing backend wiring.
Great for
- Front-end developers turning Figma frames into code that must match closely
- Designers checking whether a build faithfully follows their file
- Teams with an established design token and component library
Skip it if
- Projects without a design system or tokens, since mapping has nothing to match against and values may be guessed
- Screenshots-only work where measurements are estimates and precise matching is not reliable
- Designs that break accessibility or domain rules, which this skill defers to other skills rather than resolving
Try asking
- “Build the pricing section from this Figma link so it matches the design and our existing tokens”
- “Implement the mobile and desktop frames of this dashboard and compare the result to the designs”
- “Turn this screenshot into a React card component using our existing components and report any deviations”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 3 original files, the MIT licence and an attribution note.
build-design.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip build-design.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip build-design.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Provide a Figma link rather than a screenshot so values are exact instead of estimated
- Name the specific frames in scope to keep the diff limited to what the design shows
- Have a browser available so the agent can screenshot each width and actually verify the match
What's inside
- agents/openai.yaml
- figma.md
- SKILL.md
Redistributed unmodified from jakubkrehel/skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.