AgriAgent: Contract-Driven Planning and Capability-Aware Tool Orchestration in Real-World Agriculture
🤖 AI Summary
Existing agricultural agent systems are constrained by a unified execution paradigm, struggling to balance adaptability and robustness when handling multimodal, multi-step tasks under conditions of high task complexity heterogeneity and incomplete tooling. This work proposes AgriAgent, a two-tier hierarchical architecture wherein simple tasks are directly handled by modality-specific agents, while complex tasks are managed through a contract-driven planning mechanism that translates task specifications into capability requirements. This enables capability-aware tool orchestration and dynamic generation, supporting verifiable and recoverable multi-step execution. Experimental results demonstrate that AgriAgent significantly outperforms existing tool-centric baselines on complex agricultural tasks, achieving markedly higher execution success rates and fault tolerance.
Application Category
📝 Abstract
Intelligent agent systems in real-world agricultural scenarios must handle diverse tasks under multimodal inputs, ranging from lightweight information understanding to complex multi-step execution. However, most existing approaches rely on a unified execution paradigm, which struggles to accommodate large variations in task complexity and incomplete tool availability commonly observed in agricultural environments. To address this challenge, we propose AgriAgent, a two-level agent framework for real-world agriculture. AgriAgent adopts a hierarchical execution strategy based on task complexity: simple tasks are handled through direct reasoning by modality-specific agents, while complex tasks trigger a contract-driven planning mechanism that formulates tasks as capability requirements and performs capability-aware tool orchestration and dynamic tool generation, enabling multi-step and verifiable execution with failure recovery. Experimental results show that AgriAgent achieves higher execution success rates and robustness on complex tasks compared to existing tool-centric agent baselines that rely on unified execution paradigms. All code, data will be released at after our work be accepted to promote reproducible research.
Problem
Research questions and friction points this paper is trying to address.
agricultural agents
task complexity
tool availability
multimodal inputs
execution paradigm
Innovation
Methods, ideas, or system contributions that make the work stand out.
contract-driven planning
capability-aware tool orchestration
hierarchical agent framework
dynamic tool generation
agricultural AI
🔎 Similar Papers
No similar papers found.
