
As of November 17, 2026, Figma is giving administrators on the Organization and Enterprise plans access to an allowlist of approved OAuth applications. In practice, no third party app, including an MCP client like Cursor or Claude Code, can freely connect to the organization’s Figma account without prior approval.
Management happens from a new Extensions tab in the admin console. There, admins can see connection requests, approve or deny apps, and keep an overall view of what is already authorized. It is a mechanism close to what we already know from Slack or Google Workspace, though it arrived a bit late at Figma.
Since the arrival of Figma MCP, AI agents have been connecting directly to design files to read layers, extract CSS, or generate code. Handy, but it also opens a door: every developer who plugs in their favorite MCP client creates one more access point to the organization’s files, often without anyone on the IT side knowing about it.
With tools like Cursor, Claude Code, and other MCP compatible clients gaining momentum, this became a sensitive topic, especially for teams working on products that are still confidential (which is often the case for startups raising funds). The allowlist answers this very concrete need: knowing who has access to what, and being able to cut that access in one click.
| Before the allowlist | With the allowlist |
|---|---|
| Any OAuth app can connect freely | Only apps approved by an admin can connect |
| No centralized view of MCP connections | Dedicated Extensions tab in the admin console |
| Access discovered after the fact, often too late | Approval or denial before any connection |
At Mazette.co, we use Figma every day, in both product design and web design, and we are also starting to connect AI agents via MCP on certain projects. We had been waiting for this allowlist. Giving teams autonomy with AI tools, yes, but not at the cost of losing visibility over who is touching what. This is exactly the kind of safeguard that lets you adopt these new practices without sacrificing the security of client files.
The catch: the feature remains limited to the Organization and Enterprise plans. Smaller teams, often the growing startups we work with, do not have access to it, even though they are just as affected by the growing number of MCP clients in their workflows. We hope to see this rolled out to other plans quickly, because AI access governance is not a concern for large companies alone.
{{webdesign}}
If your organization is on an Organization or Enterprise plan, the first move is to take stock of the apps already connected before switching on the allowlist. It is better to discover existing connections upfront than to accidentally block a tool your team relies on.
If you work with complex design systems and several third party integrations, this kind of control quickly becomes essential to keep a healthy structure. It is a topic we often discuss with our clients in product design, alongside component organization and token governance.
Code
Impactful titles, polished meta descriptions, sitemaps, Schema.org, Hn structure, and audits: these are the SEO foundations that make your content readable by both Google AND generative AI. We cover both simultaneously.

With over 180 clients, our agency is one of the few in France to hold the Webflow Premium Partner label, supporting you in every area, including custom web design, sophisticated automation, conversion optimization, and SEO and GEO visibility.
We create bespoke web designs focused on performance, user journeys, and conversion. Join over 180 clients who have trusted us to elevate their brand.
