Home
AI Agents Adopt Three-Column Layout as Standard, Plugin Ecosystem Emerges as Key Battleground

Leading AI Agent products like Codex, Claude Desktop, and Cursor 3.0 have recently and almost simultaneously adopted a three-panel layout. This convergence is not mere coincidence or imitation, but a natural evolution toward the optimal interface for Agent interaction models.
The rationale is straightforward. The traditional two-column chatbot design works well for simple Q&A, but today's AI Agents can write code, edit files, and call tools autonomously. Users need a clear space to review the results of these actions, hence the dedicated workspace panel on the right. Furthermore, as users spend more time within Agents, the need to constantly switch to external tools like VSCode diminishes. The expectation now is to edit and refine outputs directly within the Agent interface. Codex's major April update highlighted this shift, transforming its right panel from a simple preview into a versatile, interactive workspace.
With this update, Codex introduced the slogan "Codex for (almost) everything," signaling ambitions beyond code generation. The goal is to support professional workflows across domains, creating a complete loop from AI generation to manual refinement. However, limitations remain for directly editing certain file formats like PPTX after generation, which may reflect a temporary strategic focus.
In the current Agent landscape, protocols like MCP effectively solve tool connectivity, while Agent Skills address the "how-to" knowledge base. Yet a significant gap remains: the need for secondary editing. No matter how capable the AI, the final 5% of precise adjustments often requires a human touch. Relying solely on Markdown editors is not a sustainable solution.
The most viable path forward points to a plugin ecosystem. The Agent would handle core orchestration, while opening file preview, secondary editing, and vertical professional capabilities to community-built plugins. Users could install these as needed to create personalized workflows. Unlike easily replicable Skills, plugins could leverage established App Store models for monetization and copyright protection, providing sustained incentives for developers. Codex has already launched an early plugin marketplace, and Cursor appears to be moving in a similar direction.
Related article
Ctrip Business Travel Unveils AI Ecosystem with 7 Industry Agents and 17 Core Capabilities
Ctrip Business Travel has officially unveiled the "Ctrip Business Travel AI Ecosystem," leveraging large language models to fundamentally restructure corporate travel management. This comprehensive ecosystem addresses four core scenarios: travel arra
Stripe to Acquire OpenRouter as AI Model Routing Gains Momentum
Stripe has agreed to acquire OpenRouter, an AI model-routing platform that gives developers access to hundreds of models through a single interface. The deal adds model selection and routing to Stripe’s existing work around AI usage and token-based b
Former OpenAI Researcher Barrett Zoph Returns to Google to Support Gemini Development
Barret Zoph, who previously abruptly left and joined OpenAI due to a disagreement with Thinking Machines Lab CEO Miriam Mularati, has once again changed employers. Alphabet's Google announced that Zoph will return as Vice President of Research, bring
Related Special Topic Recommendations
Comments (0)
0/500

Leading AI Agent products like Codex, Claude Desktop, and Cursor 3.0 have recently and almost simultaneously adopted a three-panel layout. This convergence is not mere coincidence or imitation, but a natural evolution toward the optimal interface for Agent interaction models.
The rationale is straightforward. The traditional two-column chatbot design works well for simple Q&A, but today's AI Agents can write code, edit files, and call tools autonomously. Users need a clear space to review the results of these actions, hence the dedicated workspace panel on the right. Furthermore, as users spend more time within Agents, the need to constantly switch to external tools like VSCode diminishes. The expectation now is to edit and refine outputs directly within the Agent interface. Codex's major April update highlighted this shift, transforming its right panel from a simple preview into a versatile, interactive workspace.
With this update, Codex introduced the slogan "Codex for (almost) everything," signaling ambitions beyond code generation. The goal is to support professional workflows across domains, creating a complete loop from AI generation to manual refinement. However, limitations remain for directly editing certain file formats like PPTX after generation, which may reflect a temporary strategic focus.
In the current Agent landscape, protocols like MCP effectively solve tool connectivity, while Agent Skills address the "how-to" knowledge base. Yet a significant gap remains: the need for secondary editing. No matter how capable the AI, the final 5% of precise adjustments often requires a human touch. Relying solely on Markdown editors is not a sustainable solution.
The most viable path forward points to a plugin ecosystem. The Agent would handle core orchestration, while opening file preview, secondary editing, and vertical professional capabilities to community-built plugins. Users could install these as needed to create personalized workflows. Unlike easily replicable Skills, plugins could leverage established App Store models for monetization and copyright protection, providing sustained incentives for developers. Codex has already launched an early plugin marketplace, and Cursor appears to be moving in a similar direction.
Ctrip Business Travel Unveils AI Ecosystem with 7 Industry Agents and 17 Core Capabilities
Ctrip Business Travel has officially unveiled the "Ctrip Business Travel AI Ecosystem," leveraging large language models to fundamentally restructure corporate travel management. This comprehensive ecosystem addresses four core scenarios: travel arra
Stripe to Acquire OpenRouter as AI Model Routing Gains Momentum
Stripe has agreed to acquire OpenRouter, an AI model-routing platform that gives developers access to hundreds of models through a single interface. The deal adds model selection and routing to Stripe’s existing work around AI usage and token-based b
Former OpenAI Researcher Barrett Zoph Returns to Google to Support Gemini Development
Barret Zoph, who previously abruptly left and joined OpenAI due to a disagreement with Thinking Machines Lab CEO Miriam Mularati, has once again changed employers. Alphabet's Google announced that Zoph will return as Vice President of Research, bring











