The authors create the first large-scale dataset and taxonomy of failure modes in multi-agent LLM systems to explain their limited performance gains.
hub Canonical reference
Nguyen, and Nghi D
Canonical reference. 100% of citing Pith papers cite this work as background.
13
Pith papers citing it
1
external citations · external index
Background
100% of classified citations
hub tools
citation-role summary
background 5
citation-polarity summary
roles
background 5polarities
background 5representative citing papers
SWE-EVO shows GPT-5.4 with OpenHands reaching only 25% success on complex multi-file evolution tasks versus 72.8% on SWE-Bench Verified, and introduces Fix Rate as a partial-progress metric.
A single long-horizon agent with dependency-aware traversal and RepoMemory produces more complete, consistent, hierarchical repository documentation than open and closed baselines.
AgentCollabBench shows that multi-agent reliability is limited by communication topology, with converging-DAG nodes causing synthesis bottlenecks that discard constraints and explain 7-40% of information loss variance.
Coding agents require a three-level proactivity taxonomy (Reactive, Scheduled, Situation Aware) evaluated by insight policy quality using Insight Decision Quality, Context Grounding Score, and Learning Lift.
CodeSim achieves new SOTA pass@1 scores on code generation benchmarks by using multi-agent simulation-driven planning and debugging instead of relying only on external tools.
icat-agent improves resolution rates on SWE-bench Verified and Pro by 3.6-18.5% over baselines via event-based multi-agent scaffolding and rubric-driven workflow pivoting while using the same models.
This survey categorizes agentic environments for LLMs by eight attributes and domains, introduces symbolic and neural synthesis paradigms with evaluation, and outlines four agent evolution pathways plus three environment evolution paradigms.
Agentic AI systems are shifting software engineering from line-level code generation to delegated repository-scale execution under supervision, with SWE-bench performance rising from 1.96% to 78.4% and productivity gains of 13.6-55.8%.
CodeWiki presents a unified framework for repository-level documentation across seven languages using hierarchical decomposition, recursive multi-agent processing, and multi-modal synthesis, outperforming DeepWiki by 4.73% on CodeWikiBench.
Non-linear domain-scoped parallel LLM agents achieve higher micro F1 than linear exploration and some baselines for multi-file change localization on SWE-bench Pro ansible tasks.
AssemPlanner is a ReAct-based multi-agent system that autonomously generates production plans from natural language inputs by integrating scheduling, knowledge, line balancing, and scene graph feedback.
Code Broker deploys a five-agent hierarchy that combines LLM semantic analysis with static linting to generate actionable Python code quality reports.
citing papers explorer
-
Why Do Multi-Agent LLM Systems Fail?
The authors create the first large-scale dataset and taxonomy of failure modes in multi-agent LLM systems to explain their limited performance gains.
-
SWE-EVO: Benchmarking Coding Agents in Long-Horizon Software Evolution Scenarios
SWE-EVO shows GPT-5.4 with OpenHands reaching only 25% success on complex multi-file evolution tasks versus 72.8% on SWE-Bench Verified, and introduces Fix Rate as a partial-progress metric.
-
Remember Your Trace: Memory-Guided Long-Horizon Agentic Framework for Consistent and Hierarchical Repository-Level Code Documentation
A single long-horizon agent with dependency-aware traversal and RepoMemory produces more complete, consistent, hierarchical repository documentation than open and closed baselines.
-
AgentCollabBench: Diagnosing When Good Agents Make Bad Collaborators
AgentCollabBench shows that multi-agent reliability is limited by communication topology, with converging-DAG nodes causing synthesis bottlenecks that discard constraints and explain 7-40% of information loss variance.
-
Agentic Coding Needs Proactivity, Not Just Autonomy
Coding agents require a three-level proactivity taxonomy (Reactive, Scheduled, Situation Aware) evaluated by insight policy quality using Insight Decision Quality, Context Grounding Score, and Learning Lift.
-
CODESIM: Multi-Agent Code Generation and Problem Solving through Simulation-Driven Planning and Debugging
CodeSim achieves new SOTA pass@1 scores on code generation benchmarks by using multi-agent simulation-driven planning and debugging instead of relying only on external tools.
-
Unlocking Model Potentials Through Adaptive Multi-Agent Scaffolding for Efficient Issue Resolution
icat-agent improves resolution rates on SWE-bench Verified and Pro by 3.6-18.5% over baselines via event-based multi-agent scaffolding and rubric-driven workflow pivoting while using the same models.
-
Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application
This survey categorizes agentic environments for LLMs by eight attributes and domains, introduces symbolic and neural synthesis paradigms with evaluation, and outlines four agent evolution pathways plus three environment evolution paradigms.
-
Agentic AI in the Software Development Lifecycle: Architecture, Empirical Evidence, and the Reshaping of Software Engineering
Agentic AI systems are shifting software engineering from line-level code generation to delegated repository-scale execution under supervision, with SWE-bench performance rising from 1.96% to 78.4% and productivity gains of 13.6-55.8%.
-
CodeWiki: Evaluating AI's Ability to Generate Holistic Documentation for Large-Scale Codebases
CodeWiki presents a unified framework for repository-level documentation across seven languages using hierarchical decomposition, recursive multi-agent processing, and multi-modal synthesis, outperforming DeepWiki by 4.73% on CodeWikiBench.
-
Exploration Structure in LLM Agents for Multi-File Change Localization
Non-linear domain-scoped parallel LLM agents achieve higher micro F1 than linear exploration and some baselines for multi-file change localization on SWE-bench Pro ansible tasks.
-
AssemPlanner: A Multi-Agent Based Task Planning Framework for Flexible Assembly System
AssemPlanner is a ReAct-based multi-agent system that autonomously generates production plans from natural language inputs by integrating scheduling, knowledge, line balancing, and scene graph feedback.
-
Code Broker: A Multi-Agent System for Automated Code Quality Assessment
Code Broker deploys a five-agent hierarchy that combines LLM semantic analysis with static linting to generate actionable Python code quality reports.