mlsanityData to 7 Oct 2026

Heterogeneous Agent Cohorts for Safe Open-Ended Exploration with Runtime Constraint Memory

arXiv 2607.11226 ↗First posted 13 Jul 2026cs.AIcited by 0

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

LLM agents today are caught in an awkward bind. Lock them down with static safety instructions and they rarely venture beyond the obvious; give them free reign with tools and multi-agent debate, and safety violations quickly follow. Rather than forcing a single model to juggle both creativity and caution, we separate the concerns across specialized roles. A Disrupter generates unconventional proposals, a Validator enforces hard runtime checks at the tool gateway, and a Broker pulls in distant but relevant analogies. Failures are not discarded -- they are compiled, via MCTS, into compact, signed constraint patches we call Scars. These patches are cached locally and inherited by future cohorts, turning repeated failures into reusable, low-cost runtime constraints. In a spatial-semantic sandbox (N=20 runs, p<0.01), our cohort reaches remote targets where debate fails, the Validator prevents all executed breaches, and Scars reduce token consumption by 15.1% by avoiding redundant validator checks. Furthermore, credit-based Communication Allocation Scores (CAS) restrict outbound bandwidth, reducing overall token costs by 55.9% under resource constraints.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 21 of its 36 references that are in the map. First posted on arXiv 13 Jul 2026.

Builds on

21 of 36 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024Agent-SafetyBench: Evaluating the Safety of LLM Agents19 Dec
  2. 2024AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents11 Oct

Other references · 19

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  2. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,064
  3. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,155
  4. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 2,950
  5. 2023WebArena: A Realistic Web Environment for Building Autonomous Agentscited by 1,908
  6. 2023AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversationcited by 1,906
  7. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,880
  8. 2023MetaGPT: Meta Programming for A Multi-Agent Collaborative Frameworkcited by 1,761
  9. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
  10. 2023Graph of Thoughts: Solving Elaborate Problems with Large Language Modelscited by 873
  11. 2023MemoryBank: Enhancing Large Language Models with Long-Term Memorycited by 539
  12. 2025Accelerating scientific discovery with Co-Scientistcited by 353
7 more
  1. 2023NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Railscited by 347
  2. 2026Towards End-to-End Automation of AI Researchcited by 150
  3. 2025RedDebate: Safer Responses Through Multi-Agent Red Teaming Debatescited by 7
  4. 2026The Rise of AI Agent Communities: Large-Scale Analysis of Discourse and Interaction on Moltbookcited by 6
  5. 2026Memory-Induced Tool-Drift in LLM Agentscited by 4
  6. 2026Latent Agents: A Post-Training Procedure for Internalized Multi-Agent Debatecited by 3
  7. 2026To Know is to Construct: Schema-Constrained Generation for Agent Memorycited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Governance Decay: How Context Compaction Silently Erases Safety Constraints in Long-Horizon LLM Agents
  2. 2026Toward Safe LLM Agents: A Survey of Specification, Verification, and Enforcement
  3. 2026A Policy-Driven Runtime Layer for Agentic LLM Serving
  4. 2026Safe Multi-Agent Behavior Must Be Maintained, Not Merely Asserted: Constraint Drift in LLM-Based Multi-Agent Systems
  5. 2026SiLR: Structure-Preserving Admission and Process Reward for LLM Tool Agents
  6. 2026Stale Does Not Mean Unsafe: Guard Precision for Tool-Using LLM Agents under Infrastructure State Races
  7. 2026Position: A Three-Layer Probabilistic Assume-Guarantee Architecture Is Structurally Required for Safe LLM Agent Deployment
  8. 2026When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents