Recently, the tech media TestingCatalog reported that Claude, the AI assistant from Anthropic, is testing a new feature called "Tasks Mode." This feature aims to turn Claude into a versatile intelligent agent that can provide efficient support in various complex tasks.

In the latest internal test version, Claude introduced an intuitive dual-panel layout. The right panel is used to display files uploaded by users or AI-generated context information, while the left progress area dynamically shows the execution process of multi-step tasks. This design not only allows users to clearly see the progress of the task but also helps them better understand the complex logic behind the scenes.

image.png

Different from traditional AI, when handling complex requests, Claude first generates a structured "Action Plan." If the user's instruction is not clear enough, Claude will proactively ask "Clarifying Questions" to ensure the accurate execution of the task. If the user does not respond within the specified time, Claude will also intelligently skip this step and continue with the subsequent tasks.

This new mode's flexibility allows users to intervene at any time, modify the action plan or add new context information to ensure the final result meets expectations. During execution, Claude can adjust in real time, demonstrating higher adaptability.

image.png

To enhance automation capabilities, the Tasks Mode of Claude integrates various "Skills" and "Model Context Protocols" (MCPs). Depending on the task goal, the system automatically calls necessary tools or performs code generation, web searches, and other operations. All generated outcomes, whether code snippets, articles, or web applications, are automatically archived into a dedicated "Artifact" tab for users to preview and review anytime, forming a complete workflow.