agent frameworks

Current LLM-driven autonomous agents lack scalable, secure, and maintainable architectures for tool orchestration, hindering large-scale deployment. To address this, we propose the novel “Control Plane as a Tool” paradigm, which— for the first time—abstracts tool scheduling, security policies, and extensibility mechanisms into a unified, pluggable tool interface, thereby decoupling control logic from the LLM agent core. Leveraging modular routing protocols and production-grade encapsulation, our approach significantly reduces integration complexity while enabling dynamic tool registration/removal and runtime policy updates. Empirical evaluation across diverse application scenarios demonstrates over 40% improvement in both horizontal scalability efficiency and security robustness. The architecture provides a reusable, evolution-aware foundation for controllable, embodied intelligent agents.

This study addresses the lack of systematic investigation into architectural design decisions for non-large language model components in current AI agent systems. The authors propose a protocol-guided, source code–driven empirical analysis method that enables, for the first time, transparent deconstruction of heterogeneous AI agent systems. Through cross-project qualitative coding and co-occurrence analysis of 70 open-source projects, they identify five core design dimensions—sub-agent architecture, context management, tooling systems, security mechanisms, and orchestration—and uncover their combinatorial patterns. Based on these findings, the study further distills five archetypal architectural patterns: lightweight tool-oriented, CLI framework–based, multi-agent orchestrator, enterprise system, and domain-specific vertical architectures.

The integration of large language model (LLM)-driven multi-agent systems (LMAS) into the full software engineering (SE) lifecycle remains underexplored, with no comprehensive mapping of LMAS applications across SE phases or consensus on research priorities. Method: We conduct a systematic literature review and empirical case studies using mainstream LMAS frameworks (e.g., AutoGen, CrewAI) to analyze LMAS deployment across SE stages—requirements, design, development, testing, and operations. Contribution/Results: We present the first phase-wise LMAS application taxonomy for SE, identifying critical research gaps. We propose the “SE 2.0” vision and a dual-track research agenda emphasizing *individual agent capability enhancement* and *cross-agent collaboration optimization*. Experimental evaluation demonstrates that state-of-the-art LMAS frameworks significantly improve autonomy, robustness, and scalability in real-world SE tasks—yet expose limitations in consistency, traceability, and domain-specific reasoning. This work establishes a foundational theoretical framework and actionable implementation guidelines for LLM-augmented SE.

This work addresses the fragmentation in current AI agent research stemming from the absence of a systematic architectural framework and unified evaluation standards. To bridge this gap, the paper proposes a comprehensive taxonomy encompassing components, orchestration, and deployment, offering a structured analysis of single- and multi-agent architectures, coordination mechanisms, and application scenarios. It integrates core modules—including large language models, memory systems, world models, planners, tool routers, and critic components—and synthesizes key techniques such as chain-of-thought reasoning, self-reflection, hierarchical planning, and multimodal perception. Building on this foundation, the study consolidates evaluation methodologies—spanning task suites, human preference alignment, and success rates under constraints—elucidates the sources of evaluation complexity, advocates for reproducible benchmarking practices, and highlights critical open challenges in verification, memory management, interpretability, and robustness.

This work addresses the current lack of open-source infrastructure capable of efficiently training and evaluating large-scale agents on complex tasks such as software engineering and computer operation. To this end, we propose a three-service decoupled architecture tailored for agent-environment interaction workloads, which separates the system into three independent services—model, agent, and environment—enabling fine-grained task scheduling, dynamic resource allocation, and unified interface communication. This design allows each component to scale independently and configure resources flexibly, significantly improving training efficiency and resource utilization. Experimental results demonstrate that the system can stably support tens of thousands of concurrent agent tasks, thereby filling a critical gap in infrastructure for large-scale agent training.

This study systematically investigates how tool architecture influences the behavior and performance of coding agents, holding underlying capabilities constant. Through controlled experiments on repository-scale program repair tasks, six distinct tool interfaces—ranging from bash and structured low-level APIs to natural language search, Python CodeAct, and cognitive scaffolding—are evaluated. Analysis of 11,700 agent trajectories reveals, for the first time, that the architectural design of tools—not merely their functional capacity—plays a critical role: structured low-level interfaces improve consistency across repeated attempts by 4.7×, natural language search increases access to relevant files by over 11%, and CodeAct substantially reduces both action steps (by 41.6%) and token consumption (by 56.3%), whereas cognitive scaffolding yields limited benefits.

This study addresses a critical gap in understanding how task-oriented Agent Plan artifacts in open-source software guide AI-powered coding tools. For the first time, it systematically identifies and analyzes real-world Agent Plan files from open-source projects by screening 36,710 GitHub repositories and conducting qualitative content analysis focused on Markdown-formatted planning documents. The investigation yields 85 valid Agent Plan files that span key engineering activities—including maintenance, design, and implementation—and explicitly articulate task intent while providing concrete execution steps and validation criteria. These findings reveal the instrumental role such plans play in facilitating human-AI collaborative development and underscore their practical value in structuring and communicating software engineering tasks.

Existing multi-agent collaborative systems are hindered by static workflows, sequential scheduling, and heterogeneous interfaces, leading to high complexity and poor scalability. This work proposes Agent-as-Tool, a unified paradigm that abstracts both agents and tools into a standardized, learnable action space, and introduces ParaManager—a lightweight coordinator enabling state-aware parallel subtask decomposition, delegation, and asynchronous execution. By unifying communication protocols and incorporating explicit state feedback, the framework facilitates efficient multi-agent collaboration. A two-stage training strategy—combining supervised fine-tuning with a recovery mechanism and reinforcement learning—optimizes task success rate, protocol compliance, response diversity, and reasoning efficiency. Experiments demonstrate that ParaManager achieves strong performance across multiple benchmarks and exhibits robust generalization to unseen agent pools.

This study addresses the disruptive impact of large language models and AI agent systems—capable of generating vast volumes of code—on traditional software engineering paradigms. The work proposes a new paradigm centered on agent orchestration, verification of AI-generated code, and structured human-AI collaboration. Through a structured synthesis of literature review and industry practices, it constructs a comprehensive framework encompassing education, toolchains, lifecycle management, and governance. The research reveals a fundamental shift in the nature of code—from a scarce craft artifact to a consumable commodity—and identifies the evolving role of software engineers toward system design, semantic validation, and accountability oversight. It further establishes key directions such as a verification-first software development lifecycle, offering both theoretical grounding and practical pathways for software engineering transformation in the AI era.

Agentic AI is transforming software engineering in two complementary directions: empowering development practices while simultaneously challenging traditional methodologies due to its inherent complexity, thereby necessitating a unified research agenda. This work addresses this dual challenge by organizing the Rio A2SE workshop, which brought together 18 experts from academia, industry, and research. Through structured reporting, collaborative clustering, and consensus-building techniques, the initiative establishes the first community-driven framework encompassing both “agent-augmented software engineering” and “software engineering for agents.” The resulting framework systematically identifies six core research areas—governance, agent software engineering, architecture, quality assurance, sustainability, and code—and articulates short- and long-term priorities, offering a foundational agenda to foster coordinated innovation across academic and industrial communities.