# Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails

arXiv: https://arxiv.org/abs/2609.09134 · HTML: https://mlsanity.com/paper/2609.09134/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 13 of its 15 references that are in the map.
- First posted on arXiv 8 Sep 2026.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 13 of 13)

- [GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning](https://mlsanity.com/paper/2507.19457/) (2025)
- [Meta-Harness: End-to-End Optimization of Model Harnesses](https://mlsanity.com/paper/2603.28052/) (2026)
- [SKILL0: In-Context Agentic Reinforcement Learning for Skill Internalization](https://mlsanity.com/paper/2604.02268/) (2026)
- [HarnessX: A Composable, Adaptive, and Evolvable Agent Harness Foundry](https://mlsanity.com/paper/2606.14249/) (2026)
- [Adapting the Interface, Not the Model: Runtime Harness Adaptation for Deterministic LLM Agents](https://mlsanity.com/paper/2605.22166/) (2026)
- [SIA: Self Improving AI with Harness & Weight Updates](https://mlsanity.com/paper/2605.27276/) (2026)
- [HarnessForge: Joint Harness and Policy Evolution for Adaptive Agent Systems](https://mlsanity.com/paper/2606.01779/) (2026)
- [Imitation Learning for Multi-turn LM Agents via On-policy Expert Corrections](https://mlsanity.com/paper/2512.14895/) (2025)
- [Co-Harness: Co-Evolving Harnesses and Model Weights for LLM Agents](https://mlsanity.com/paper/2607.22688/) (2026)
- [Trajectory2Task: Training Robust Tool-Calling Agents with Synthesized Yet Verifiable Data for Complex User Intents](https://mlsanity.com/paper/2601.20144/) (2026)
- [Better Harnesses, Smaller Models: Building 90% Cheaper Agents via Automated Harness Adaptation](https://mlsanity.com/paper/2607.08938/) (2026)
- [Compiling Agentic Workflows into LLM Weights: Near-Frontier Quality at Two Orders of Magnitude Less Cost](https://mlsanity.com/paper/2605.22502/) (2026)
- [A Few Teacher Steps Go a Long Way: Cost-Efficient On-Policy Data Augmentation for Agent Post-Training](https://mlsanity.com/paper/2607.04574/) (2026)

## Built on by (most cited first, 4 of 4)

- [Scaling Trajectories for Complex Tasks through Recursive Self-Rewrite](https://mlsanity.com/paper/2610.02826/) (2026)
- [RLHarness: Co-evolving Procedural Skills with Reinforcement Learning for Long-horizon Multimodal Reasoning](https://mlsanity.com/paper/2609.32326/) (2026)
- [EvoIn: Bridging Evolution and Internalization for Agent Fine-Tuning](https://mlsanity.com/paper/2609.35290/) (2026)
- [Harness Evolution Hits a Ceiling: When Weight Training Should Begin](https://mlsanity.com/paper/2610.11655/) (2026)

## Abstract (arXiv)

Agent harnesses (the system prompt, tool set, execution hooks, and context-management scaffolding around a model) are a critical determinant of agentic task success. Automated harness evolution can enable smaller models to perform well on domain-specific tasks at a fraction of frontier-model cost. Since both the harness and model weights shape behavior, we ask how harness evolution and lightweight fine-tuning should be combined. Across seven enterprise agent tasks, we first evolve a harness with the weaker model, then find that a stronger expert often uses it more effectively, suggesting expert supervision could close the remaining gap. However, training the weaker model on the expert's complete trajectories under the evolved harness backfires: performance regresses on all seven tasks by 4 to 30 points across Qwen3-Coder and Gemma 4, even though the same procedure helps under the unevolved harness. Our analysis shows that imitation transfers knowledge and increases scaffold usage, but disrupts model-harness fit: the weaker model adopts the expert's planning strategy without the competence to execute it and no longer matches the harness evolved around its native planning style. We therefore develop an on-policy expert-correction pipeline, automated by a meta-level MLE agent, that localizes the failing turn in the weaker model's own rollout and asks the expert to rewrite only that turn. This preserves the model's planning style and combines the gains of harness evolution and model adaptation. Our results identify and resolve a source of contention between harness and weight updates, yielding a compatibility-preserving recipe for economical co-evolution on domain-specific enterprise tasks.