Skill Detail

Read and modify Figma designs from coding agents with Talk to Figma MCP

Bridge Cursor, Claude Code, and other MCP clients into Figma so agents can inspect selections, create nodes, annotate designs, and apply bulk edits.

Integrations & ConnectorsMCP
Integrations & Connectors MCP Security Reviewed
⭐ 6.8k GitHub stars ⬇ 11.1k/wk npm
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill read-and-modify-figma-designs-from-coding-agents-with-talk-to-figma-mcp Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
Bun, MCP-compatible client, Figma desktop or web app, Talk to Figma Figma plugin
Install & setup
Install Bun, run `bun setup`, start the bridge with `bun socket`, install the Talk to Figma plugin from Figma Community or locally, then connect the plugin channel and register the MCP server in the agent client.
Author
Grab
Publisher
Organization
Last updated
Jun 13, 2026
Quick brief

Use Talk to Figma MCP when a coding or design agent needs to work directly with an open Figma file, not just read exported layout context. The operator runs the WebSocket bridge, connects the MCP server to the agent client, opens the Figma plugin, joins a channel, and then asks the agent to inspect selections, create elements, add annotations, or propagate repeated design edits. The boundary is live Figma read/write automation through MCP and the companion plugin, not a generic Figma API wrapper or design-product listing.

How it works

What this skill actually does

Inputs and prerequisites: Bun, MCP-compatible client, Figma desktop or web app, Talk to Figma Figma plugin.

Setup notes: Install Bun, run `bun setup`, start the bridge with `bun socket`, install the Talk to Figma plugin from Figma Community or locally, then connect the plugin channel and register the MCP server in the agent client.

Source and verification boundary: use https://www.figma.com/community/plugin/1485687494525374295/cursor-talk-to-figma-mcp-plugin as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.

Framework fit: publish this as a MCP workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.