GoKawiilGoogle Research releases Harness² framework for self-improving AI agent systems
Google Research has released Harness², an open-source framework that lets AI agents refine their own operating 'harness'—prompts, tools, memory and workflow—without changing the underlying model's weights. The system works through two feedback loops: one that tweaks the harness for the task at hand by comparing different execution attempts, and another that keeps a persistent record of lessons learned to guide future tasks. Tests on GLM-5 and Gemini 3.5 Flash showed performance gains of up to 11.5 points on one benchmark and 11.8 percentage points on another after a single refinement step.
GoKawiil's interpretation of the reporting above, not reported fact.
By improving performance through better scaffolding rather than retraining, Harness² suggests a cheaper route to boosting agent capability on new, unseen tasks. The persistent 'harnessing-level' memory could let agent systems accumulate practical know-how over time, potentially narrowing the gap between general-purpose models and specialized, hand-tuned agent setups.
- Harness² improves agent performance by editing the surrounding system, not the model weights
- Gains of up to 11.5 points and 11.8 percentage points were measured across professional benchmarks
- The framework supports multiple executors and benchmarks, including JobBench, LAB, and WorkBuddy-Bench
Source: github.com, 2026-10-19
Published there as: “Harness2 by Google Research – Recursive Agent Harnessing”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.