Agent Harness Distillation: Inference-Time Harness Extraction and Exploitation in Autonomous Multi-Agent Systems

arXiv:2607.28147 · cs.CR · Submitted 2026-08-24 · Read on arXiv

Yu Cui, Wuli Yang, Yirui Shi, Junhao Xia, Hui Jiang, Lei Gao, Chenfu Bao

cs.CR

Submitted: 2026-08-24

Updated: 2026-08-25

Comments: This work is currently in progress

Code: https://github.com/openai/codex

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Distillation — 59× in this paper · explained in 2 episodes
• Agent Harness — 14× in this paper · explained in 2 episodes
• Agent Harness Distillation — 12× in this paper
• Multi-Agent System — 11× in this paper · explained in 6 episodes
• Knowledge Distillation — 8× in this paper · explained in 11 episodes
• Accuracy — 7× in this paper · explained in 1 episode
• Autonomous Multi-Agent Systems — 5× in this paper
• Benchmark — 4× in this paper · explained in 2 episodes
• Fidelity — 4× in this paper · explained in 2 episodes
• Alignment — 3× in this paper · explained in 1 episode
• Generalization — 3× in this paper · explained in 13 episodes
• Information Leakage — 3× in this paper · explained in 2 episodes
• Trace — 3× in this paper · explained in 2 episodes
• Verification — 3× in this paper · explained in 2 episodes
• Adaptation — 2× in this paper · explained in 1 episode
• Core — 2× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 2× in this paper · explained in 6 episodes
• Self-Distillation — 2× in this paper · explained in 7 episodes
• Abstraction — 1× in this paper · explained in 1 episode
• Adversarial Attacks — 1× in this paper · explained in 1 episode
• Agentic AI — 1× in this paper · explained in 1 episode
• Agentic AI systems — 1× in this paper · explained in 1 episode
• Agentic Coding — 1× in this paper · explained in 1 episode
• Aggregation — 1× in this paper · explained in 1 episode
• Benchmarking — 1× in this paper · explained in 3 episodes
• Clean Transfer — 1× in this paper
• Clean Visibility — 1× in this paper
• Injected Fidelity — 1× in this paper
• Inverse Problem — 1× in this paper · explained in 2 episodes
• Probe Perturbation — 1× in this paper
• Steering — 1× in this paper · explained in 2 episodes
• Two-Stage Framework — 1× in this paper · explained in 2 episodes

Sources

• MemoHarness: Agent Harnesses That Learn from Experience
• Recursive Harness Self-Improvement
• FlowSteer: Prompt-Only Workflow Steering Exposes Planning-Time Vulnerabilities in Multi-Agent LLM Systems
• DeepSeek-V3 Technical Report
• HarnessForge: Joint Harness and Policy Evolution for Adaptive Agent Systems
• ClawKeeper: Comprehensive Safety Protection for OpenClaw Agents Through Skills, Plugins, and Watchers
• HarnessX: A Composable, Adaptive, and Evolvable Agent Harness Foundry
• Harness Handbook: Making Evolving Agent Harnesses Readable,Navigable, and Editable
• Mathematical Reasoning in Large Language Models: Assessing Logical and Arithmetic Errors across Wide Numerical Ranges
• Adapting the Interface, Not the Model: Runtime Harness Adaptation for Deterministic LLM Agents
• Analyzing Defensive Misdirection Against Model-Guided Automated Attacks on Agentic AI Systems
• Qwen3 Technical Report
• Training with Harnesses: On-Policy Harness Self-Distillation for Complex Reasoning

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs