mlsanityData to 7 Oct 2026

Interactive Debugging and Steering of Multi-Agent AI Systems

arXiv 2503.02068 ↗First posted 3 Mar 2025cs.MA cs.AI cs.HCcited by 68

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

Abstract (arXiv)

Fully autonomous teams of LLM-powered AI agents are emerging that collaborate to perform complex tasks for users. What challenges do developers face when trying to build and debug these AI agent teams? In formative interviews with five AI agent developers, we identify core challenges: difficulty reviewing long agent conversations to localize errors, lack of support in current tools for interactive debugging, and the need for tool support to iterate on agent configuration. Based on these needs, we developed an interactive multi-agent debugging tool, AGDebugger, with a UI for browsing and sending messages, the ability to edit and reset prior agent messages, and an overview visualization for navigating complex message histories. In a two-part user study with 14 participants, we identify common user strategies for steering agents and highlight the importance of interactive message resets for debugging. Our studies deepen understanding of interfaces for debugging increasingly important agentic workflows.

Key facts. Cited by 68 papers in the map (data through 7 Oct 2026). Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use. Builds on 14 of its 22 references that are in the map. First posted on arXiv 3 Mar 2025.

Builds on

14 of 22 references.

References · 14

  1. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,383
  2. 2017Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPRcited by 1,086
  3. 2020On Faithfulness and Factuality in Abstractive Summarizationcited by 895
  4. 2024Large Language Model based Multi-Agents: A Survey of Progress and Challengescited by 752
  5. 2024WebVoyager: Building an End-to-End Web Agent with Large Multimodal Modelscited by 424
  6. 2021AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Promptscited by 265
  7. 2021Polyjuice: Generating Counterfactuals for Explaining, Evaluating, and Improving Modelscited by 219
  8. 2024Who Validates the Validators? Aligning LLM-Assisted Evaluation of LLM Outputs with Human Preferencescited by 142
  9. 2022PromptChainer: Chaining Large Language Model Prompts through Visual Programmingcited by 134
  10. 2023ChainForge: A Visual Toolkit for Prompt Engineering and LLM Hypothesis Testingcited by 88
  11. 2023EvalLM: Interactive Evaluation of Large Language Model Prompts on User-Defined Criteriacited by 73
  12. 2023Zeno: An Interactive Framework for Behavioral Evaluation of Machine Learningcited by 33
2 more
  1. 2023ConstitutionMaker: Interactively Critiquing Large Language Models by Converting Feedback into Principlescited by 33
  2. 2016On Human Intellect and Machine Failures: Troubleshooting Integrative Machine Learning Systemscited by 18

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 29; 2026: 39.

68 papers.

Sort:

By area:

  1. 2025Why Do Multi-Agent LLM Systems Fail?cited by 559
  2. 2025TRAIL: Trace Reasoning and Agentic Issue Localizationcited by 98
  3. 2025Understanding Software Engineering Agents: A Study of Thought-Action-Result Trajectoriescited by 75
  4. 2025MARFT: Multi-Agent Reinforcement Fine-Tuningcited by 51
  5. 2025Who is Introducing the Failure? Automatically Attributing Failures of Multi-Agent Systems via Spectrum Analysiscited by 33
  6. 2025ARE: Scaling Up Agent Environments and Evaluationscited by 28
  7. 2026Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environmentscited by 28
  8. 2025DoVer: Intervention-Driven Auto Debugging for LLM Multi-Agent Systemscited by 23
  9. 2025RAFFLES: Reasoning-based Attribution of Faults for LLM Systemscited by 19
  10. 2025Automatic Failure Attribution and Critical Step Prediction Method for Multi-Agent Systems Based on Causal Inferencecited by 18
  11. 2025A Comprehensive Survey on Benchmarks and Solutions in Software Engineering of LLM-Empowered Agentic Systemcited by 15
  12. 2025CORRECT: COndensed eRror RECognition via knowledge Transfer in multi-agent systemscited by 14
38 more
  1. 2025Towards Transparent and Incentive-Compatible Collaboration in Decentralized LLM Multi-Agent Systems: A Blockchain-Driven Approachcited by 11
  2. 2026TrajAudit: Automated Failure Diagnosis for Agentic Coding Systemscited by 11
  3. 2025Demystifying the Lifecycle of Failures in Platform-Orchestrated Agentic Workflowscited by 10
  4. 2025Traceability and Accountability in Role-Specialized Multi-Agent LLM Pipelinescited by 10
  5. 2026DiLLS: Interactive Diagnosis of LLM-based Multi-agent Systems via Layered Summary of Agent Behaviorscited by 7
  6. 2025Agentic Services Computingcited by 6
  7. 2025When Should Users Check? Modeling Confirmation Frequency in Multi-Step Agentic AI Taskscited by 6
  8. 2026When Agents Fail: A Comprehensive Study of Bugs in LLM Agents with Automated Labelingcited by 6
  9. 2026Human oversight of agentic systems in practice: Examining the oversight work, challenges, and heuristics of developers using software agentscited by 6
  10. 2026MAS-PromptBench: When Does Prompt Optimization Improve Multi-Agent LLM Systems?cited by 4
  11. 2026AgentCheck: A Reproduce-Intervene-Mitigate Workbench for LLM Agents over MCPcited by 4
  12. 2025RAG Without the Lag: Interactive Debugging for Retrieval-Augmented Generation Pipelinescited by 3
  13. 2025On the Regulatory Potential of User Interfaces for AI Agent Governancecited by 3
  14. 2025XAgen: An Explainability Tool for Identifying and Correcting Failures in Multi-Agent Workflowscited by 3
  15. 2026The Behavioral Fabric of LLM-Powered GUI Agents: Human Values and Interaction Outcomescited by 3
  16. 2026Are We All Using Agents the Same Way? An Empirical Study of Core and Peripheral Developers Use of Coding Agentscited by 3
  17. 2026Overseeing Agents Without Constant Oversight: Challenges and Opportunitiescited by 3
  18. 2026When Planning Fails Despite Correct Execution: On Epistemic Calibration for LLM-Based Multi-Agent Systemscited by 3
  19. 2026AI Agents Push Humans Out of the Loopcited by 3
  20. 2025Understanding User Experiences of Computer Use Agents: Design Space and Opportunities for Building Agent UX Prototypescited by 2
  21. 2026More Than "Means to an End": Supporting Reasoning with Transparently Designed AI Data Science Processescited by 2
  22. 2026How to Steer Your Multi-Agent System: Human-LLM Collaborative Planningcited by 2
  23. 2026Knowledge-Based Zero-Replay Debugging of Multi-Agent LLM Tracescited by 2
  24. 2026TraceView: Interactive Visualization of Agentic Program Repair Trajectoriescited by 2
  25. 2026MafiaScope: Non-Invasive, Time-Resolved Belief Probing for LLM Agents in Social Deduction Gamescited by 2
  26. 2026AppWorld-UL: Benchmarking Diverse Agent-User Interactions for Tool-Usecited by 2
  27. 2025AgentDynEx: Nudging the Mechanics and Dynamics of Multi-Agent Simulationscited by 1
  28. 2025ORCA: ORchestrating Causal Agentcited by 1
  29. 2025DoubleAgents: Human-Agent Alignment in a Socially Embedded Workflowcited by 1
  30. 2025Botender: Supporting Communities in Collaboratively Designing AI Agents through Case-Based Provocationscited by 1
  31. 2025Interactive Training: Feedback-Driven Neural Network Optimizationcited by 1
  32. 2026"When to Hand Off, When to Work Together": Understanding Concurrent Human-Agent Interaction in Shared Co-Creative Workspacescited by 1
  33. 2026InconLens: Interactive Visual Diagnosis of Behavioral Inconsistencies in LLM-based Agentic Systemscited by 1
  34. 2026QiMeng-PRepair: Precise Code Repair via Edit-Aware Reward Optimizationcited by 1
  35. 2026SelfHeal: Empirical Fix Pattern Analysis and Bug Repair in LLM Agentscited by 1
  36. 2026HANSEL: Extracting Breadcrumbs from Web Agent Trajectories for Interactive Verificationcited by 1
  37. 2026SAFARI: Scaling Long Horizon Agentic Fault Attribution via Active Investigationcited by 1
  38. 2026Autonomous Repair for Multi-Agent Systems via Monte-Carlo Tree Searchcited by 1

50 most cited of 68 All 68 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024AutoGen Studio: A No-Code Developer Tool for Building and Debugging Multi-Agent Systems
  2. 2026AgentStepper: Interactive Debugging of Software Development Agents
  3. 2024UniDebugger: Hierarchical Multi-Agent Framework for Unified Software Debugging
  4. 2026Empowering Autonomous Debugging Agents with Efficient Dynamic Analysis
  5. 2025AgentCompass: Towards Reliable Evaluation of Agentic Workflows in Production
  6. 2026AgentDebugX: An Open-Source Toolkit for Failure Observability, Attribution, and Recovery in LLM Agents
  7. 2026Debug2Fix: Can Interactive Debugging Help Coding Agents Fix More Bugs?
  8. 2024Debug Smarter, Not Harder: AI Agents for Error Resolution in Computational Notebooks