OpenSage: Self-programming Agent Generation Engine
🤖 AI Summary
Existing agent development kits (ADKs) rely on manually designed topologies, toolsets, and memory mechanisms, which limit agents’ generalization and performance. This work proposes the first large language model–driven engine for the automatic generation of complete AI agents, capable of autonomously constructing agent architectures and toolsets while introducing a structured hierarchical graph memory system and a dynamic sub-agent management mechanism. The framework also integrates a domain-specific toolkit tailored for software engineering tasks. By enabling AI to autonomously generate full-fledged agents, this approach marks a paradigm shift from human-centered to AI-centered development. Evaluated on three mainstream benchmarks, the method significantly outperforms existing ADKs, and ablation studies confirm the effectiveness of each component, demonstrating superior generalization and task performance.
Application Category
📝 Abstract
Agent development kits (ADKs) provide effective platforms and tooling for constructing agents, and their designs are critical to the constructed agents' performance, especially the functionality for agent topology, tools, and memory. However, current ADKs either lack sufficient functional support or rely on humans to manually design these components, limiting agents' generalizability and overall performance. We propose OpenSage, the first ADK that enables LLMs to automatically create agents with self-generated topology and toolsets while providing comprehensive and structured memory support. OpenSage offers effective functionality for agents to create and manage their own sub-agents and toolkits. It also features a hierarchical, graph-based memory system for efficient management and a specialized toolkit tailored to software engineering tasks. Extensive experiments across three state-of-the-art benchmarks with various backbone models demonstrate the advantages of OpenSage over existing ADKs. We also conduct rigorous ablation studies to demonstrate the effectiveness of our design for each component. We believe OpenSage can pave the way for the next generation of agent development, shifting the focus from human-centered to AI-centered paradigms.
Problem
Research questions and friction points this paper is trying to address.
Agent Development Kits
agent topology
toolsets
memory support
generalizability
Innovation
Methods, ideas, or system contributions that make the work stand out.
self-programming agents
agent development kit
hierarchical graph memory
automatic toolset generation
LLM-based agent architecture
🔎 Similar Papers
No similar papers found.
