mlsanityData to 14 Oct 2026

RobustFlow: Towards Robust Agentic Workflow Generation

arXiv 2509.21834 ↗First posted 26 Sep 2025cs.MAcited by 20

On the map: LLM agents and tool use → Software engineering coding agents

Abstract (arXiv)

The automated generation of agentic workflows is a promising frontier for enabling large language models (LLMs) to solve complex tasks. However, the empirical study reveals that existing agentic workflow generation methods are not robust. They often generate inconsistent workflows when provided with instructions that are semantically equivalent but phrased differently. This brittleness severely undermines their reliability in real-world applications. To tackle this challenge, we propose RobustFlow, a robust agentic workflow generation system that leverages preference optimization to learn invariance across instruction variations. We also introduce a benchmark of semantically equivalent instruction variants with node-level and graph-level metrics for evaluating workflow generation robustness. By training on these instruction variants, RobustFlow achieves workflow generation robustness scores of 70%--90% across diverse perturbations, outperforming existing approaches including AFlow and ScoreFlow. The code is publicly available at https://github.com/DEFENSE-SEU/RobustFlow.

Key facts. Cited by 20 papers in the map (data through 14 Oct 2026). Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use. Builds on 44 of its 69 references that are in the map. First posted on arXiv 26 Sep 2025.

Builds on

44 of 69 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming29 May
  2. 2025GNNs as Predictors of Agentic Workflow Performances14 Mar
  3. 2025ScoreFlow: Mastering LLM Agent Workflows via Score-based Preference Optimization6 Feb
  4. 2025Flow: Modularized Agentic Workflow Automation14 Jan
  5. 2024AFlow: Automating Agentic Workflow Generation14 Oct
  6. 2024Benchmarking Agentic Workflow Generation10 Oct
  7. 2024AutoFlow: Automated Workflow Generation for Large Language Model Agents1 Jul
  8. 2023ChatDev: Communicative Agents for Software Development16 Jul

Other references · 36

  1. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  2. 2019Sentence-BERT: Sentence Embeddings using Siamese BERT-Networkscited by 9,170
  3. 2025Qwen3 Technical Reportcited by 8,820
  4. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  5. 2014Explaining and Harnessing Adversarial Examplescited by 7,364
  6. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  7. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  8. 2024GPT-4o System Cardcited by 4,588
  9. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  10. 2021Program Synthesis with Large Language Modelscited by 3,637
  11. 2023A Survey on Large Language Model based Autonomous Agentscited by 2,242
  12. 2023MetaGPT: Meta Programming for A Multi-Agent Collaborative Frameworkcited by 1,808
24 more
  1. 2019DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphscited by 903
  2. 2023AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviorscited by 598
  3. 2024JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Modelscited by 536
  4. 2024Automated Design of Agentic Systemscited by 407
  5. 2024SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuningcited by 355
  6. 2024TextGrad: Automatic "Differentiation" via Textcited by 273
  7. 2023PromptRobust: Towards Evaluating the Robustness of Large Language Models on Adversarial Promptscited by 226
  8. 2024Data Interpreter: An LLM Agent For Data Sciencecited by 225
  9. 2023AutoAgents: A Framework for Automatic Agent Generationcited by 216
  10. 2025Multi-agent Architecture Search via Agentic Supernetcited by 179
  11. 2024AgentSquare: Automatic LLM Agent Search in Modular Design Spacecited by 155
  12. 2024The Good, The Bad, and The Greedy: Evaluation of LLMs Should Not Ignore Non-Determinismcited by 116
  13. 2023T-Eval: Evaluating the Tool Utilization Capability of Large Language Models Step by Stepcited by 93
  14. 2020InfoBERT: Improving Robustness of Language Models from An Information Theoretic Perspectivecited by 92
  15. 2024Non-Determinism of "Deterministic" LLM Settingscited by 92
  16. 2021Evaluating the Robustness of Neural Language Models to Input Perturbationscited by 87
  17. 2024CLLMs: Consistency Large Language Modelscited by 72
  18. 2024Robust Preference Optimization through Reward Model Distillationcited by 67
  19. 2025FlowReasoner: Reinforcing Query-Level Meta-Agentscited by 63
  20. 2024Improving the Robustness of Large Language Models via Consistency Alignmentcited by 36
  21. 2024Contrastive Instruction Tuningcited by 17
  22. 2023VerifAI: Verified Generative AIcited by 16
  23. 2024Large Language Models are Contrastive Reasonerscited by 11
  24. 2024SelfPrompt: Autonomously Evaluating LLM Robustness via Domain-Constrained Knowledge Guidelines and Refined Adversarial Promptscited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 18.

20 papers.

Sort:

By area:

  1. 2026ReCreate: Reasoning and Creating Domain Agents Driven by Experiencecited by 24
  2. 2026From Static Templates to Dynamic Runtime Graphs: A Survey of Workflow Optimization for LLM Agentscited by 24
  3. 2026Harnessing Agentic Evolutioncited by 11
  4. 2026FlowSteer: Towards Agents Designing Agentic Workflows via Reinforced Progressive Canvas Editingcited by 9
  5. 2026RemoteAgent: Bridging Vague Human Intents and Earth Observation with RL-based Agentic MLLMscited by 8
  6. 2026Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Applicationcited by 8
  7. 2025CoS: Towards Optimal Event Scheduling via Chain-of-Schedulingcited by 4
  8. 2026Learning to Compose for Cross-domain Agentic Workflow Generationcited by 3
  9. 2026Do We Always Need Query-Level Workflows? Rethinking Agentic Workflow Generation for Multi-Agent Systemscited by 2
  10. 2026PeroMAS: A Multi-agent System of Perovskite Material Discoverycited by 2
  11. 2026RemoteShield: Enable Robust Multimodal Large Language Models for Earth Observationcited by 2
  12. 2026Towards Direct Evaluation of Harness Optimizers via Priority Rankingcited by 2
8 more
  1. 2025GLOW: Graph-Language Co-Encoding for Agentic Workflow Performance Predictioncited by 1
  2. 2026Automating the Design of Embodied Agent Architecturescited by 1
  3. 2026FlowEvo: Self-Evolving Agents through the Co-Evolution of Workflows and Executable Skillscited by 1
  4. 2026Coupled Hierarchical Search over Topology and Execution for Agentic Workflow Synthesiscited by 1
  5. 2026Generating Workflow DAGs from Natural Language with Non-Reasoning LLMscited by 0
  6. 2026Robust Workflow Generation via Adversarial Learning for Audio Deepfake Detectioncited by 0
  7. 2026MoFlow: Multi-Objective Agentic Workflow Generationcited by 0
  8. 2026CB-Orchestrator: Adaptive Workflow Optimization for LLM Agents via Contextual Banditscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026It Takes Workflows to Evolve Better Workflows
  2. 2026GenOS: Compositional Certificates for Semantic Robustness in AI Code Generation
  3. 2026From Search to Synthesis: Training LLMs as Zero-Shot Workflow Generators
  4. 2025AutoSynth: Automated Workflow Optimization for High-Quality Synthetic Dataset Generation via Monte Carlo Tree Search
  5. 2026Workflow-Localized Mechanism Learning: Attribution-Guided Repair and Knowledge Reuse for Structured Agent Skills
  6. 2025A2Flow: Automating Agentic Workflow Generation via Self-Adaptive Abstraction Operators
  7. 2025AdaptFlow: Adaptive Workflow Optimization via Meta-Learning
  8. 2025SEW: Self-Evolving Agentic Workflows for Automated Code Generation