agentsclimarketplace

Figmahandoff

Skill saurabhshuklagrowisto/saurabh-ai-systems/claude-skills/figmahandoff

AI architect for GTM and martech. I design and ship production agentic systems for B2B sales and marketing: lead scraping and scoring with an eval gated learning loop, autonomous CRM enrichment, ABM pipelines, live dashboards that refresh themselves, MCP servers and Claude skills. Built at Growisto.

Install
npx -y skills add saurabhshuklagrowisto/saurabh-ai-systems --skill figmahandoff

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use this skill EVERY TIME a Figma URL is mentioned or pasted in any message. Triggers on any figma.com link, any mention of "figma", or any request to implement/build/code a design. Before writing any code or reading any design, this skill intercepts the request and asks two required questions: (1) which tech stack — React or Shopify Liquid, and (2) whether to show the Figma design system first or proceed directly. Do NOT skip these questions. Do NOT write any code before both are answered. This is mandatory for all Figma-related implementation tasks at Acme.

SKILL.md

5.5 KB, as published. Nobody here has run it

Figma Handoff Skill

Purpose

Every time a developer shares a Figma URL, Claude must stop and ask two questions before doing anything else. No code. No design reading. No assumptions. Questions first.

This ensures:

  • Code is generated in the right tech stack (React vs Shopify Liquid)
  • Developer can inspect the design system before committing to implementation
  • Output is consistent across all engineers on the team

Trigger

This skill activates when:

  • A message contains a figma.com URL
  • A message contains the word "figma" alongside any implementation intent ("build", "implement", "code", "create", "make")
  • A developer says "here is the design" or "use this design" or similar

Step 1 — Ask the Two Required Questions

The moment a Figma URL is detected, respond with exactly this — no code, no design analysis yet:


Got the Figma link. Before I start, two quick questions:

1. Which tech stack should I use?

  • A — React
  • B — Shopify Liquid

2. How would you like to proceed?

  • C — Show me the Figma design system first (components, tokens, styles)
  • D — Go straight to implementation

Reply with your choices (e.g. "A, D" or "B, C") and I'll get started.


Do not deviate from this format. Do not add extra questions. Wait for the developer's reply before doing anything with the Figma URL.


Step 2 — Handle the Response

If they chose C (Show Design System First)

Read the Figma file using the MCP Server and display a full design system breakdown in this structure:

## Design System — [Frame/File Name]

### Components
List every named component Claude can identify from the file:
- ComponentName — description of what it is / how it's used

### Colour Tokens
- token-name: #hexvalue — usage note

### Typography
- style-name: Font Family, Size, Weight, Line Height — usage note

### Spacing & Layout
- spacing scale values
- border radius values
- shadow / elevation values

### Notes
Anything unusual, missing, or worth flagging before implementation starts

After displaying this, ask:

"Which components do you want me to implement? You can name them or say 'all'."

Wait for confirmation before writing any code.


If they chose D (Go Straight to Implementation)

Proceed directly to reading the Figma URL and generating code.


Step 3 — Generate Code Based on Tech Stack

If React (A)

  • Read the full React patterns reference before generating: references/react.md
  • Generate a functional React component using TypeScript if the codebase uses it
  • Use Tailwind CSS utility classes for styling — map Figma values to Tailwind scale (see reference file)
  • Check tailwind.config.js for project colour tokens before hardcoding any hex values
  • Import existing components from the codebase — do not reinvent them (check src/components/ui/ first)
  • Use cn() utility for conditional or merged class names
  • Use cva for components with multiple Figma variants
  • If Code Connect is set up, use mapped component imports automatically

If Shopify Liquid (B)

  • Generate valid Liquid template syntax (.liquid file)
  • Use Shopify section schema where relevant ({% schema %})
  • Reference Shopify theme settings variables ({{ settings.color_background }}) over hardcoded values
  • Use {{ 'style.css' | asset_url | stylesheet_tag }} pattern for styles
  • Keep logic in Liquid, styles in a companion .css file unless told otherwise
  • Read the references file for detailed Shopify Liquid patterns: references/shopify-liquid.md

Step 4 — After Code Generation

Always end with:

## Implementation Notes
- [List any design tokens that were hardcoded because they weren't in the Figma variables]
- [List any components that were created from scratch because Code Connect isn't set up]
- [List any design decisions made due to ambiguity in the Figma file]

## To improve accuracy
- Set up Code Connect to map Figma components to your codebase (see the Figma MCP Evaluation doc)
- Share the frame URL for any specific sub-component you want refined

Rules — Never Break These

  1. Never write code before both questions are answered
  2. Never assume React — always ask, even if it looks like a React project
  3. Never hardcode colours or spacing if design tokens exist in the Figma file
  4. Never skip the Design System view if the developer asks for it — show everything Claude can read
  5. Never write to the Figma file — read only, always (write-to-canvas is off until the team agrees a protocol)

Reference Files

  • references/react.md — React patterns, TypeScript interfaces, Tailwind token mapping, layout patterns, common mistakes
    • Load this when tech stack answer is A (React)
  • references/shopify-liquid.md — Shopify Liquid patterns, schema structure, theme settings usage
    • Load this when tech stack answer is B (Shopify Liquid)

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.