What the Figma Agent Can Do After Config 2026
Figma used its Config 2026 conference on June 24 to significantly expand what its AI agent can do, pushing it well beyond a simple prompt-and-generate tool.
The most practical addition is Agent Skills. Teams can now create repeatable commands for the Figma agent, turning frequently used workflows into something the whole team can run again without having to remember how it was done the first time. Users write those skills in plain text prompts, with no coding required.
Context is also getting richer. The Figma agent can now connect to outside services including Notion, GitHub, and Microsoft Excel, and users can bring in additional context through file attachments and web search as well.
Frequently used workflows can be saved as skills and deployed team-wide, with one example being a design team checking a layout against regulatory compliance standards.
The Figma agent also now handles shader generation. Users describe the visual effect they want, whether that is frosted glass, dithering, or pixelation, and the agent generates it along with adjustable controls, removing the need for graphics programming knowledge.
On the plugin side, users can describe what they want, such as a layout generator or a vector path tracer, and Figma will build the plugin. The feature extends the platform’s existing plugin ecosystem by lowering the barrier from writing code to writing a sentence.
The agent is also coming to more surfaces, including FigJam and Slides, with early access available via the waitlist at figma.com/config-betas.
The Figma agent is currently free during the beta period, with AI credits not consumed for any prompts or actions. Once it becomes generally available, standard AI credit usage will apply.
