Here are answers to some questions you may have while working with the Figma MCP server.
General questions
What are skills?
Markdown files that teach agents specific Figma workflows: the steps, the order, and the conventions. Users should download the foundational Figma skill first, then create or download other skills to use alongside it. Anyone can author and share new ones. No code, no approval process.
Where can I see example skills?
You can see the skills created here on the community: figma.com/community/skills
Write to canvas
Note: We're quickly improving how Figma supports AI agents. This will eventually be a usage-based paid feature, but is currently available for free during the beta period.
What is use_figma?
With the use_figma tool in Figma's MCP server, AI agents can now write directly to your Figma files — creating and modifying real design assets using your existing components, variables, and tokens. Agents read your library first and build with what's already there, so the output reflects your design system.
What's the difference between the use_figma tool and the figma-use skill?
- use_figma is the tool that allows agents to write to Figma files through our MCP server. It executes Plugin API code that creates and modifies real Figma objects — components, variables, frames, auto layout — with full awareness of your existing design system.
- /figma-use is a skill — the foundational set of instructions that teaches agents how to properly use the use_figma tool. Other skills build on top of it. You need both.
Why can't I find the use_figma tool?
Make sure you're using a supported client connected to the remote Figma MCP server, and that your client supports skills (specifically write to canvas).
If you're unsure whether you're using the remote server, you can ask your agent. For example, "Which version of the Figma MCP server am I using?"
If you've confirmed that you're using the remote Figma MCP server and you're still not seeing the tools, you may need to reinstall the MCP server, or reauthenticate to it.
Why can't I see the Figma skills?
If you can't find the figma-use skill, make sure:
- You're using a supported client.
- If your client supports plugins, make sure you've installed the Figma plugin. Otherwise, make sure you've downloaded the skills.
- Make sure the skills are in the correct file location for your MCP client. The exact location depends on your client. You can ask your agent to move them to the right spot. For example, "I downloaded mcp-server-guide-main.zip. Move the agent skills in the zip to the right place."
Why aren't I getting the results I want?
Make sure you have the /figma-use foundational skill installed — it gives agents a shared understanding of how Figma works and significantly improves output quality. You can also try additional skills for specific workflows, or write your own to match how your team works. Remember this is a beta — designs generated by agents may require manual updates and changes.
In many MCP clients, you can directly invoke the skill: /figma-use <your prompt here>
You can also tell your agent explicitly to use the skill: "Use figma-use to <your prompt here>"
Is my client supported?
All of our supported clients are currently listed here in the Guide to the Figma MCP server.
How is this different from the code to canvas feature?
The features are different tools for different jobs.
Code to Canvas (generate_figma_design) captures a running app as flat Figma layers — fast mockups, but no library connection.
Write to canvas (use_figma) is structured generation — agents use your existing components and variables to produce real design system assets. They complement each other. For example, you might use Code to Canvas to bring the latest designs from code into Figma, then use use_figma to refine them with your existing components and variables.
Do I need a Full Seat to use use_figma's write functionality?
A Full Seat is required to write to Figma files outside of drafts. Dev Seats get read-only access outside of their drafts — design context, screenshots, metadata, variable definitions, and use-figma read APIs. Both are useful workflows.
Is this safe to use on production files?
You can, but because the tool is in beta and still being actively developed, we would recommend:
- Testing on a duplicate of your library or a draft file to ensure you are happy with the outputs prior to altering your production files
- Keeping a duplicate file or local copy as a backup
What are the known limitations for writing to canvas?
- 20kb output response limit per call
- No assets (image) support yet - including no importing components that have images/videos, or creating GIFs
- Custom fonts aren't supported yet
- Components must be manually published before Code Connect completes
- Beta-level quality — output may need manual review and cleanup
- The tool is evolving and will improve through the beta period
Why am I seeing a gray area / box instead of my images?
We don't support images in the use_figma tool yet, so you may be seeing placeholders generated instead of the image, or a component with assets won't have the asset included - but this is coming soon!