Atomic Task Graph: A Unified Framework for Agentic Planning and Execution

2026-07-02

AI Summary

Existing large language model agents tackling multi-step tasks often rely on costly recomputation or task-specific fine-tuning, resulting in poor generalization and limited reusability of intermediate results. This work proposes the Atomic Task Graph (ATG) framework, which— for the first time—explicitly models task decomposition and execution dependencies using a unified directed acyclic graph. During planning, ATG recursively decomposes high-level tasks; during execution, it enables parallel scheduling and local backtracking for error recovery. Notably, ATG operates effectively across diverse tasks without any training, achieving significant performance gains over strong baselines on three interactive benchmarks using only lightweight 7B–8B parameter models, while simultaneously improving both task success rates and execution efficiency.

Abstract

LLM-based agents have shown strong potential for solving complex multi-step tasks, yet existing performance improvements often rely on either scaling to larger backbone models or task-specific fine-tuning. The former incurs substantial computational costs, while the latter typically generalizes poorly across different tasks. Although prompt-based control is training-free and broadly applicable, existing methods still leave input-output dependencies between subtasks implicit in textual trajectories, making verified intermediate results difficult to reuse. To address these limitations, we propose Atomic Task Graph (ATG), a unified control framework for planning and execution. Specifically, ATG maintains an explicit graph to expose dependencies and support reuse. During planning, it recursively decomposes a high-level task into subtasks, forming a sequence of directed acyclic graphs (DAGs) whose evolution can be traced. During execution, the dependencies exposed by ATG allow independent branches to be executed in parallel, thereby improving execution efficiency. When failures are detected, ATG leverages the graph evolution history to localize the error source and repair only the affected region, preserving validated regions unchanged. Experiments show that ATG consistently outperforms strong baselines in success rate and execution efficiency across three interactive benchmarks using only 7B-8B backbones.

Problem

• LLM-based agents
• multi-step tasks
• subtask dependencies
• intermediate result reuse
• generalization

Innovation

• Atomic Task Graph
• Agentic Planning
• Task Decomposition
• Dependency Graph
• Execution Efficiency