EE-MCP: Self-Evolving MCP-GUI Agents via Automated Environment Generation and Experience Learning
For developers of computer-use agents, this work provides a principled framework and automated pipeline to balance GUI and API modalities, enabling self-improvement without manual intervention.
The paper formulates MCP-GUI interplay as a hybrid policy learning problem and proposes a self-evolving framework with automatic environment generation, trajectory collection, and an experience bank for inference-time improvement. On MCP-dominant tasks, distillation achieves 77.8% pass rate (+17.8pp), while the experience bank improves GUI-intensive tasks by +10.0pp.
Computer-use agents that combine GUI interaction with structured API calls via the Model Context Protocol (MCP) show promise for automating software tasks. However, existing approaches lack a principled understanding of how agents should balance these two modalities and how to enable iterative self-improvement across diverse applications. We formulate MCP-GUI interplay as a unified hybrid policy learning problem where the agent learns when each modality provides complementary advantages, and show that distillation and experience augmentation target fundamentally different failure modes - requiring application-aware mechanism selection. Built on this formulation, we propose a self-evolving framework with a fully automatic pipeline that orchestrates automatic environment generation and validation, trajectory collection, gap-driven task synthesis, and quality-filtered training - all without manual intervention. A key innovation is our experience bank, which accumulates LLM-learned rules from trajectory comparison, enabling inference-time improvement without fine-tuning. Systematic \textbf{cross-application analysis} across three desktop applications reveals that the optimal strategy depends on MCP-GUI composition: distillation achieves 77.8\% pass rate on MCP-dominant tasks (+17.8pp), while the experience bank excels on GUI-intensive tasks (+10.0pp).