Figma Co-founder Dylan Field Exposes MCP Protocol’s Critical Flaws

Key Takeaways:

• Figma is significantly expanding its support for the Model Context Protocol (MCP), an open-source standard for AI agents, despite prior reservations from co-founder Dylan Field.
• Field’s concerns center on MCP’s perceived “asymmetric context sharing” and potential advantages for “frontier labs,” questioning its true “openness” when certain entities don’t fully participate.
• The company balances its commitment to an open ecosystem with a cautious approach, citing risks of data abuse as a primary reason for not fully exposing all Figma data via MCP.

Figma’s AI Paradox: Embracing Open Standards While Questioning Their Full Openness

In a move that highlights the complex dance between fostering open ecosystems and safeguarding proprietary data, Figma, the leading design collaboration platform, has announced a significant expansion of its support for the Model Context Protocol (MCP). Unveiled at its inaugural Config conference in India, this initiative brings 20 new partners – including industry heavyweights like Linear, Notion, and Mistral – into the MCP fold, standardizing how AI agents connect to Figma data and tools.

Yet, this embrace comes with a layer of palpable tension, openly acknowledged by Figma co-founder Dylan Field. His public misgivings about MCP’s true “openness” have cast an intriguing shadow over what would otherwise be a straightforward narrative of industry collaboration. Field’s skepticism points to a broader debate simmering within the tech community: how genuinely open are “open standards” when the biggest players might not be playing by the same rules?

The Co-Founder’s Critique: Asymmetry and the Frontier Labs

Field’s concerns aren’t new, nor are they abstract. Days before the Config India announcement, the co-creator of MCP voiced frustration over Figma’s perceived lack of support for an MCP-enabled tool called Pi. Field’s response on X, the social media platform, pulled back the curtain on his deeper anxieties, specifically calling out Anthropic, one of MCP’s creators.

Can you help me understand why you don’t make various Anthropic resources available via MCP? Maybe I’m missing something.

— Dylan Field (@zoink) September 30, 2026

“I find it interesting that Anthropic (creators of MCP standard) do not expose artifacts / other cloud resources created via MCP,” Field wrote. He elaborated on what he sees as a critical imbalance: “The debate is framed as open vs closed. Everyone always prefers open… obviously. And meanwhile issues of asymmetric context sharing are ignored.” (Field later clarified that Figma does plan to support Pi, mitigating the immediate concern).

During a subsequent media briefing, I pressed Field on this ongoing debate. He articulated a nuanced position: while Figma is genuinely committed to fostering an open ecosystem, he doesn’t view MCP as a flawless solution. “I’m really excited about what we’re doing in terms of opening up our ecosystem,” Field stated. “I also think that MCP is not a perfect protocol. And I think that it is something that has tons of room for improvement.”

Field further suggested that the current iteration of MCP might be “advantaged more to frontier labs”—a term often used to describe large, well-resourced AI research companies. This implies a fear that the protocol, while ostensibly open, could inadvertently benefit those with the most data and computational power, potentially creating an uneven playing field for smaller developers and startups.

The Balancing Act: Openness, Utility, and Preventing Abuse

Figma’s CEO didn’t shy away from the practical implications of full openness. He clarified that the company’s measured approach stems from a fundamental concern: the more data Figma exposes through MCP, the higher the potential for misuse. While acknowledging that the vast majority of users have no malicious intent, the company bears the responsibility of considering and mitigating those risks.

This stance highlights a critical dilemma facing all platform providers in the age of ubiquitous AI agents. As AI systems become more sophisticated and capable of interacting with vast datasets, the question of data governance, security, and ethical use moves to the forefront. Opening up APIs and data streams for AI agents promises immense utility and innovation, but it also creates new vectors for data breaches, privacy violations, competitive intelligence gathering, or even the creation of harmful AI applications.

Figma’s strategy appears to be one of cautious optimism: embracing MCP to empower its users and integrate with the burgeoning AI agent landscape, while simultaneously advocating for improvements to the protocol itself to ensure genuine equity and security. This involves working with partners to define what responsible data sharing looks like within an open standard, rather than a blanket exposure of all resources.

The Broader Implications for AI Standards

Figma’s public deliberation over MCP underscores a vital evolutionary phase for AI agent standards. Protocols like MCP are crucial for realizing the vision of autonomous AI agents that can seamlessly interact with diverse applications and services. They aim to break down data silos, enabling agents to retrieve information, execute commands, and collaborate across different platforms – a key step towards truly intelligent, interconnected AI ecosystems.

However, as Field’s comments suggest, the path to truly equitable and secure open standards is fraught with challenges. Debates around governance, transparency, and power dynamics will continue to shape how these protocols evolve. Figma, with its influential position in the design and product development world, is not just adopting a standard; it’s actively participating in the conversation about how that standard should be built, governed, and improved for the benefit of all, not just a select few.

The Bottom Line

Figma’s expansion of MCP support, juxtaposed with Dylan Field’s candid critiques, paints a vivid picture of the tech industry’s complex journey into the AI-powered future. It’s a testament to the powerful tension between the desire for open innovation and the imperative of responsible data stewardship. As AI agents become more integral to our digital lives, companies will continue to navigate this delicate balance, pushing for standards that foster interoperability while demanding protocols robust enough to prevent abuse and ensure a level playing field for all participants. Figma’s measured approach may well become a template for how other major platforms engage with emerging AI standards, shaping an ecosystem that is both open and secure.