pith. machine review for the scientific record. sign in

arxiv: 2402.03578 · v3 · submitted 2024-02-05 · 💻 cs.MA · cs.AI

Recognition: 2 theorem links

· Lean Theorem

LLM Multi-Agent Systems: Challenges and Open Problems

Shanshan Han , Qifan Zhang , Weizhao Jin , Zhaozhuo Xu
Authors on Pith no claims yet

Pith reviewed 2026-05-17 10:18 UTC · model grok-4.3

classification 💻 cs.MA cs.AI
keywords multi-agent systemslarge language modelstask allocationiterative reasoningcontext managementmemory managementblockchain applicationsagent collaboration
0
0 comments X

The pith

Multi-agent LLM systems can solve complex tasks through agent collaboration but leave several challenges inadequately addressed.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper explores how systems of multiple large language model agents can divide labor and interact to tackle problems beyond the reach of any single agent. It focuses on five areas where progress has been limited: deciding which agent handles which part of a task, using repeated debates to strengthen reasoning, organizing layered context across agents, maintaining useful memory during long interactions, and applying these systems inside blockchain environments. By mapping these gaps the authors intend to steer future work toward practical fixes that would make multi-agent setups more reliable in real distributed applications.

Core claim

By leveraging the diverse capabilities and roles of individual agents, multi-agent systems can tackle complex tasks through agent collaboration, yet several challenges remain inadequately addressed including task allocation, robust reasoning through iterative debates, complex context management, memory management, and blockchain applications.

What carries the argument

Agent collaboration through role specialization and iterative interaction, which allows the system to distribute work and refine outputs across multiple LLM instances.

If this is right

  • Better task allocation methods would allow agents to divide work more efficiently and reduce redundant computation.
  • Iterative debate protocols could produce more reliable final answers by letting agents challenge one another's outputs.
  • Improved context layering would let systems maintain coherence across longer, multi-turn conversations involving many agents.
  • Stronger memory mechanisms would support persistent state across sessions, enabling agents to build on prior joint work.
  • Blockchain integrations could extend multi-agent coordination to decentralized settings where trust and record-keeping matter.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Solutions developed for these multi-agent challenges may also improve single-agent LLM performance when applied to internal self-critique or planning loops.
  • Empirical tests that isolate the effect of fixing one challenge at a time would clarify which gaps are most costly in practice.
  • The listed challenges overlap with known issues in distributed computing, suggesting cross-field techniques could be adapted.

Load-bearing premise

That the specific challenges of task allocation, iterative debate reasoning, context and memory management, and blockchain use are currently not being handled adequately enough to guide future progress without further systematic study.

What would settle it

A new benchmark or survey that demonstrates existing multi-agent LLM implementations already achieve high performance on complex tasks without needing major improvements in any of the listed challenge areas.

read the original abstract

This paper explores multi-agent systems and identify challenges that remain inadequately addressed. By leveraging the diverse capabilities and roles of individual agents, multi-agent systems can tackle complex tasks through agent collaboration. We discuss optimizing task allocation, fostering robust reasoning through iterative debates, managing complex and layered context information, and enhancing memory management to support the intricate interactions within multi-agent systems. We also explore potential applications of multi-agent systems in blockchain systems to shed light on their future development and application in real-world distributed systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper surveys LLM-based multi-agent systems, arguing that agent collaboration enables solutions to complex tasks and that several challenges remain inadequately addressed: optimizing task allocation, achieving robust reasoning via iterative debates, managing layered context information, improving memory handling for multi-agent interactions, and applying such systems to blockchain for real-world distributed environments. The work positions itself as identifying open problems to guide future development.

Significance. If the identified challenges are shown to be genuine gaps, the paper could help direct research attention toward practical issues in scaling LLM multi-agent systems, particularly the underexplored blockchain angle. Its qualitative framing offers a high-level map rather than new methods or data, so impact would depend on whether the discussion adds novel synthesis beyond existing surveys.

major comments (1)
  1. [Abstract] Abstract and opening discussion of challenges: the repeated claim that task allocation, iterative debate reasoning, context/memory management, and blockchain integration 'remain inadequately addressed' is presented as a premise without a literature-gap analysis, citation counts, or concrete examples of shortcomings in prior systems. This assertion is load-bearing for the paper's framing as a survey of open problems.
minor comments (2)
  1. The manuscript would be strengthened by adding a summary table that lists each challenge alongside representative prior works and their documented limitations.
  2. Notation for agent roles and context layers is introduced informally; explicit definitions or a diagram would improve readability for readers outside the immediate subfield.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback on our survey of LLM-based multi-agent systems. We have revised the manuscript to strengthen the substantiation of the identified open challenges.

read point-by-point responses

  1. Referee: [Abstract] Abstract and opening discussion of challenges: the repeated claim that task allocation, iterative debate reasoning, context/memory management, and blockchain integration 'remain inadequately addressed' is presented as a premise without a literature-gap analysis, citation counts, or concrete examples of shortcomings in prior systems. This assertion is load-bearing for the paper's framing as a survey of open problems.

    Authors: We agree that explicitly supporting the premise with gap analysis strengthens the framing. In the revised manuscript, we have added a new subsection early in the introduction that summarizes our literature review process across recent multi-agent LLM papers. This includes references to prior surveys, notes on the relatively sparse coverage of certain topics (supported by our review counts), and concrete examples of limitations such as suboptimal task decomposition in frameworks like AutoGen and inconsistent outcomes in debate-based reasoning setups. These changes directly address the load-bearing claim while preserving the paper's focus on identifying open problems. revision: yes

Circularity Check

0 steps flagged

No circularity: survey lists open problems without derivations or self-referential reductions

full rationale

The paper is a high-level survey identifying challenges in LLM multi-agent systems (task allocation, iterative debates, context/memory management, blockchain applications). It makes no claims of first-principles derivations, fitted parameters, predictions, or uniqueness theorems. The central positioning—that listed issues remain inadequately addressed—rests on assertion rather than quantitative gap analysis, but this is a weakness in evidence quality, not a circular reduction of any result to its own inputs by construction. No equations, self-citations as load-bearing premises, or renamings of known results appear. The document is self-contained as an opinionated overview of open questions.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The paper is a survey of open problems and introduces no new mathematical parameters, axioms, or postulated entities; all content rests on standard domain knowledge of LLM and multi-agent systems.

pith-pipeline@v0.9.0 · 5375 in / 1058 out tokens · 103928 ms · 2026-05-17T10:18:34.277186+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 18 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Why Do Multi-Agent LLM Systems Fail?

    cs.AI 2025-03 unverdicted novelty 8.0

    The authors create the first large-scale dataset and taxonomy of failure modes in multi-agent LLM systems to explain their limited performance gains.

  2. MASPrism: Lightweight Failure Attribution for Multi-Agent Systems Using Prefill-Stage Signals

    cs.SE 2026-05 unverdicted novelty 7.0

    MASPrism attributes failures in LLM multi-agent executions by extracting token-level negative log-likelihood and attention weights from a small model's prefill pass, then ranking candidates with a second prefill, achi...

  3. MASPrism: Lightweight Failure Attribution for Multi-Agent Systems Using Prefill-Stage Signals

    cs.SE 2026-05 unverdicted novelty 7.0

    MASPrism attributes failures in multi-agent systems by ranking candidates from prefill-stage NLL and attention signals of a 0.6B SLM, beating baselines by up to 33.41% Top-1 accuracy and proprietary LLMs by up to 89.5...

  4. Dr.Sai: An agentic AI for real-world physics analysis at BESIII

    hep-ex 2026-04 unverdicted novelty 7.0

    Dr.Sai autonomously executed full physics analysis pipelines on real BESIII data to re-measure ten J/psi decay branching fractions, matching established benchmarks without any manual coding.

  5. Weak-Link Optimization for Multi-Agent Reasoning and Collaboration

    cs.AI 2026-04 unverdicted novelty 7.0

    WORC improves multi-agent LLM reasoning to 82.2% average accuracy by predicting and compensating for the weakest agent via targeted extra sampling rather than uniform reinforcement.

  6. ACIArena: Toward Unified Evaluation for Agent Cascading Injection

    cs.AI 2026-04 unverdicted novelty 7.0

    ACIArena provides a unified specification, attack suites across external inputs, profiles, and messages, plus 1,356 test cases over six MAS implementations, demonstrating that topology alone is insufficient for robust...

  7. FuzzAgent: Multi-Agent System for Evolutionary Library Fuzzing

    cs.SE 2026-05 conditional novelty 6.0

    FuzzAgent deploys specialized agents that collaborate on harness generation, execution, and crash triage to evolve fuzzing campaigns, delivering 45-191% more branch coverage than four baselines on 20 C/C++ libraries a...

  8. LLM-X: A Scalable Negotiation-Oriented Exchange for Communication Among Personal LLM Agents

    cs.AI 2026-05 unverdicted novelty 6.0

    LLM-X is a scalable architecture for direct negotiation and communication among personal LLM agents, featuring federated gateways, typed protocols, and policy enforcement, shown stable in experiments with up to 12 agents.

  9. Sustaining Cooperation in Populations Guided by AI: A Folk Theorem for LLMs

    cs.GT 2026-05 unverdicted novelty 6.0

    A folk theorem for LLMs proves that all feasible and individually rational outcomes can be sustained as ε-equilibria in repeated games where LLMs advise client populations, despite indirect observation.

  10. Explicit Trait Inference for Multi-Agent Coordination

    cs.AI 2026-04 unverdicted novelty 6.0

    ETI lets LLM agents infer and track partners' psychological traits (warmth and competence) from histories, cutting payoff loss 45-77% in games and boosting performance 3-29% on MultiAgentBench versus CoT baselines.

  11. CoopGuard: Stateful Cooperative Agents Safeguarding LLMs Against Evolving Multi-Round Attacks

    cs.CR 2026-04 unverdicted novelty 6.0

    CoopGuard deploys cooperative agents to track conversation history and counter evolving multi-round attacks on LLMs, achieving a 78.9% reduction in attack success rate on a new 5,200-sample benchmark.

  12. AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents

    cs.AI 2025-03 unverdicted novelty 6.0

    AgentSpec introduces a customizable DSL for runtime enforcement of safety constraints on LLM agents, achieving over 90% prevention of unsafe code actions, zero hazardous embodied actions, and 100% AV compliance in eva...

  13. Beyond Individual Intelligence: Surveying Collaboration, Failure Attribution, and Self-Evolution in LLM-based Multi-Agent Systems

    cs.AI 2026-05 conditional novelty 5.0

    The survey proposes the LIFE framework to unify fragmented research on collaboration, failure attribution, and self-evolution in LLM multi-agent systems into a progression toward self-organizing intelligence.

  14. A Comprehensive Survey of Self-Evolving AI Agents: A New Paradigm Bridging Foundation Models and Lifelong Agentic Systems

    cs.AI 2025-08 unverdicted novelty 5.0

    A comprehensive review of self-evolving AI agents that improve themselves over time, organized via a framework of inputs, agent system, environment, and optimizers, with domain-specific and safety discussions.

  15. AssemPlanner: A Multi-Agent Based Task Planning Framework for Flexible Assembly System

    cs.RO 2026-05 unverdicted novelty 4.0

    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.

  16. Agentic Microphysics: A Manifesto for Generative AI Safety

    cs.CY 2026-04 unverdicted novelty 4.0

    The authors introduce agentic microphysics and generative safety to link local agent interactions to population-level risks in agentic AI through a causally explicit framework.

  17. Multi-Agent Collaboration Mechanisms: A Survey of LLMs

    cs.AI 2025-01 unverdicted novelty 4.0

    The survey organizes LLM-based multi-agent collaboration mechanisms into a framework with dimensions of actors, types, structures, strategies, and coordination protocols, reviews applications across domains, and ident...

  18. LLM-Based Multi-Agent Systems for Code Generation: A Multi-Vocal Literature Review

    cs.SE 2026-02 unverdicted novelty 3.0

    A review of 114 studies classifies motivations into nine categories, analyzes common models and benchmarks, synthesizes challenges into six categories with 26 subcategories and solutions, and identifies six future res...

Reference graph

Works this paper leans on

50 extracted references · 50 canonical work pages · cited by 17 Pith papers · 15 internal anchors

  1. [1]

    arXiv preprint arXiv:2311.11855 , year=

    Evil geniuses: Delving into the safety of llm-based agents , author=. arXiv preprint arXiv:2311.11855 , year=

    work page arXiv

  2. [2]

    Identifying the Risks of LM Agents with an LM-Emulated Sandbox

    Identifying the risks of lm agents with an lm-emulated sandbox , author=. arXiv preprint arXiv:2309.15817 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  3. [3]

    arXiv preprint arXiv:2311.11797 , year=

    Igniting Language Intelligence: The Hitchhiker's Guide From Chain-of-Thought Reasoning to Language Agents , author=. arXiv preprint arXiv:2311.11797 , year=

    work page arXiv

  4. [4]

    arXiv preprint arXiv:2401.10019 , year=

    R-Judge: Benchmarking Safety Risk Awareness for LLM Agents , author=. arXiv preprint arXiv:2401.10019 , year=

    work page arXiv

  5. [5]

    Multi-Agent Security Workshop@ NeurIPS'23 , year=

    I See You! Robust Measurement of Adversarial Behavior , author=. Multi-Agent Security Workshop@ NeurIPS'23 , year=

    work page

  6. [6]

    The International Journal of High Performance Computing Applications , volume=

    Data-driven scalable pipeline using national agent-based models for real-time pandemic response and decision support , author=. The International Journal of High Performance Computing Applications , volume=. 2023 , publisher=

    work page 2023

  7. [7]

    Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security

    Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security , author=. arXiv preprint arXiv:2401.05459 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  8. [8]

    Tree of Thoughts: Deliberate Problem Solving with Large Language Models

    Tree of thoughts: Deliberate problem solving with large language models , author=. arXiv preprint arXiv:2305.10601 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  9. [9]

    arXiv preprint arXiv:2305.08291 , year=

    Large Language Model Guided Tree-of-Thought , author=. arXiv preprint arXiv:2305.08291 , year=

    work page arXiv

  10. [10]

    Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology , pages=

    Generative agents: Interactive simulacra of human behavior , author=. Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology , pages=

    work page

  11. [11]

    Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing , pages=

    Mot: Memory-of-thought enables chatgpt to self-improve , author=. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing , pages=

    work page 2023

  12. [12]

    arXiv preprint arXiv:2306.07174 , year=

    Augmenting Language Models with Long-Term Memory , author=. arXiv preprint arXiv:2306.07174 , year=

    work page arXiv

  13. [13]

    arXiv preprint arXiv:2305.17653 , year=

    Prompt-Guided Retrieval Augmentation for Non-Knowledge-Intensive Tasks , author=. arXiv preprint arXiv:2305.17653 , year=

    work page arXiv

  14. [14]

    arXiv preprint arXiv:2207.00747 , year=

    Rationale-augmented ensembles in language models , author=. arXiv preprint arXiv:2207.00747 , year=

    work page arXiv

  15. [15]

    2024 , publisher=

    A Causal Framework for AI Regulation and Auditing , author=. 2024 , publisher=

    work page 2024

  16. [16]

    Advances in Neural Information Processing Systems , volume=

    Retrieval-augmented generation for knowledge-intensive nlp tasks , author=. Advances in Neural Information Processing Systems , volume=

    work page

  17. [17]

    Multi-Agent Security Workshop@ NeurIPS'23 , year=

    Language Agents as Hackers: Evaluating Cybersecurity Skills with Capture the Flag , author=. Multi-Agent Security Workshop@ NeurIPS'23 , year=

    work page

  18. [18]

    Advances in Neural Information Processing Systems , volume=

    Webshop: Towards scalable real-world web interaction with grounded language agents , author=. Advances in Neural Information Processing Systems , volume=

    work page

  19. [19]

    Reflexion: Language Agents with Verbal Reinforcement Learning

    Reflexion: an autonomous agent with dynamic memory and self-reflection , author=. arXiv preprint arXiv:2303.11366 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  20. [20]

    arXiv preprint arXiv:2305.17390 , year=

    Swiftsage: A generative agent with fast and slow thinking for complex interactive tasks , author=. arXiv preprint arXiv:2305.17390 , year=

    work page arXiv

  21. [21]

    Gorilla: Large Language Model Connected with Massive APIs

    Gorilla: Large language model connected with massive apis , author=. arXiv preprint arXiv:2305.15334 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  22. [22]

    WebArena: A Realistic Web Environment for Building Autonomous Agents

    Webarena: A realistic web environment for building autonomous agents , author=. arXiv preprint arXiv:2307.13854 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  23. [23]

    arXiv preprint arXiv:2305.17126 , year=

    Large language models as tool makers , author=. arXiv preprint arXiv:2305.17126 , year=

    work page arXiv

  24. [24]

    International Conference on Machine Learning , pages=

    Oracles & followers: Stackelberg equilibria in deep multi-agent reinforcement learning , author=. International Conference on Machine Learning , pages=. 2023 , organization=

    work page 2023

  25. [25]

    Multi-Agent Security Workshop@ NeurIPS'23 , year=

    Second-order Jailbreaks: Generative Agents Successfully Manipulate Through an Intermediary , author=. Multi-Agent Security Workshop@ NeurIPS'23 , year=

    work page

  26. [26]

    Electronics , volume=

    Emergent Cooperation and Strategy Adaptation in Multi-Agent Systems: An Extended Coevolutionary Theory with LLMs , author=. Electronics , volume=. 2023 , publisher=

    work page 2023

  27. [27]

    2010 , publisher=

    Market structure and equilibrium , author=. 2010 , publisher=

    work page 2010

  28. [28]

    Proceedings of the 7th ACM conference on Electronic commerce , pages=

    Computing the optimal strategy to commit to , author=. Proceedings of the 7th ACM conference on Electronic commerce , pages=

    work page

  29. [29]

    Multi-Agent Security Workshop@ NeurIPS'23 , year=

    Stackelberg Games with Side Information , author=. Multi-Agent Security Workshop@ NeurIPS'23 , year=

    work page

  30. [30]

    Improving Factuality and Reasoning in Language Models through Multiagent Debate

    Improving Factuality and Reasoning in Language Models through Multiagent Debate , author=. arXiv preprint arXiv:2305.14325 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  31. [31]

    Show Your Work: Scratchpads for Intermediate Computation with Language Models

    Show your work: Scratchpads for intermediate computation with language models , author=. arXiv preprint arXiv:2112.00114 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  32. [32]

    Advances in Neural Information Processing Systems , volume=

    Chain-of-thought prompting elicits reasoning in large language models , author=. Advances in Neural Information Processing Systems , volume=

    work page

  33. [33]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Multi-agent discussion mechanism for natural language generation , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  34. [34]

    Artificial Intelligence Review , pages=

    Multi-agent deep reinforcement learning: a survey , author=. Artificial Intelligence Review , pages=. 2022 , publisher=

    work page 2022

  35. [35]

    Game Theory , pages=

    Nash equilibrium , author=. Game Theory , pages=. 1989 , publisher=

    work page 1989

  36. [36]

    Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks

    Program of thoughts prompting: Disentangling computation from reasoning for numerical reasoning tasks , author=. arXiv preprint arXiv:2211.12588 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  37. [37]

    Tab- C o T : Zero-shot Tabular Chain of Thought

    Ziqi, Jin and Lu, Wei. Tab- C o T : Zero-shot Tabular Chain of Thought. Findings of the Association for Computational Linguistics: ACL 2023. 2023. doi:10.18653/v1/2023.findings-acl.651

    work page doi:10.18653/v1/2023.findings-acl.651 2023

  38. [38]

    arXiv preprint arXiv:2308.09687 , year=

    Graph of thoughts: Solving elaborate problems with large language models , author=. arXiv preprint arXiv:2308.09687 , year=

    work page arXiv

  39. [39]

    Self-Consistency Improves Chain of Thought Reasoning in Language Models

    Self-consistency improves chain of thought reasoning in language models , author=. arXiv preprint arXiv:2203.11171 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  40. [40]

    Proceedings of the 2016 ACM SIGSAC conference on computer and communications security , pages=

    On the security and performance of proof of work blockchains , author=. Proceedings of the 2016 ACM SIGSAC conference on computer and communications security , pages=

    work page 2016

  41. [41]

    The Review of financial studies , volume=

    Blockchain without waste: Proof-of-stake , author=. The Review of financial studies , volume=. 2021 , publisher=

    work page 2021

  42. [42]

    Ethereum project yellow paper , volume=

    Ethereum: A secure decentralised generalised transaction ledger , author=. Ethereum project yellow paper , volume=

    work page

  43. [43]

    Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate

    Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate , author=. arXiv preprint arXiv:2305.19118 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  44. [44]

    Feudal Multi-Agent Hierarchies for Cooperative Reinforcement Learning

    Feudal multi-agent hierarchies for cooperative reinforcement learning , author=. arXiv preprint arXiv:1901.08492 , year=

    work page internal anchor Pith review Pith/arXiv arXiv 1901

  45. [45]

    ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debate

    Chateval: Towards better llm-based evaluators through multi-agent debate , author=. arXiv preprint arXiv:2308.07201 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  46. [46]

    Multi-Agent Collaboration: Harnessing the Power of Intelligent LLM Agents

    Multi-Agent Collaboration: Harnessing the Power of Intelligent LLM Agents , author=. arXiv preprint arXiv:2306.03314 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  47. [47]

    2023 , eprint=

    Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View , author=. 2023 , eprint=

    work page 2023

  48. [48]

    CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society

    Camel: Communicative agents for" mind" exploration of large scale language model society , author=. arXiv preprint arXiv:2303.17760 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  49. [49]

    arXiv preprint arXiv:2309.15943 , year=

    Scalable Multi-Robot Collaboration with Large Language Models: Centralized or Decentralized Systems? , author=. arXiv preprint arXiv:2309.15943 , year=

    work page arXiv

  50. [50]

    arXiv preprint arXiv:2308.12503 , year=

    Cgmi: Configurable general multi-agent interaction framework , author=. arXiv preprint arXiv:2308.12503 , year=

    work page arXiv