El Agente Gráfico：科学智能体的结构化执行图

Large language models (LLMs) are increasingly utilized to automate scientific workflows, yet their integration with heterogeneous computational tools remains ad hoc and fragile. Current agentic approaches often depend on unstructured text for managing context and coordinating execution, resulting in overwhelming volumes of information that can obscure decision provenance and hinder auditability. This research presents El Agente Gráfico, a single-agent framework that embeds LLM-driven decision-making within a type-safe execution environment and dynamic knowledge graphs for external persistence. Central to this approach is a structured abstraction of scientific concepts and an object-graph mapper that represents computational state as typed Python objects.