{"slug": "harness-2-recursive-agent-harnessing-for-an-open-world", "title": "Harness^2: Recursive Agent Harnessing for an Open World", "summary": "A new arXiv paper (2610.14224v1) introduces Harness^2, a plug-and-play recursive agent harnessing framework that improves LLM agent performance without ground-truth verifiers. Across three professional benchmarks and four model-harness configurations, a single recursion step improved GLM-5 and Gemini 3.5 Flash over their base harnesses by up to 11.5 points on WorkBuddy-Bench, 4.5 points on JobBench, and 11.8 percentage points on LAB, with further gains as the task-level recursion budget increases. The framework combines task-level recursion, which uses a contrastive Propose-Probe-Compose step to guide harness edits, with harnessing-level recursion that extracts experience to update a continual improvement harness for future tasks.", "body_md": "arXiv:2610.14224v1 Announce Type: new \nAbstract: LLM agents' capabilities depend not only on their underlying models but also on their execution harnesses. Recursive self-improvement (RSI) offers a promising approach to automatically evolve these harnesses. However, deploying RSI in open-world settings still faces two critical challenges. First, open-world tasks are typically unseen during the harness design phase and novel at test time, leaving no ground-truth verifiers to reliably guide online harness edits. Second, the RSI process remains highly task-specific and fails to generalize to novel scenarios. We introduce Harness^2, a plug-and-play, recursive agent harnessing framework that addresses these challenges through two connected levels of recursion. Task-level recursion refines the current task's harness using a contrastive Propose-Probe-Compose step that reliably guides edits without ground-truth verifiers, and can be scaled further via parallel or sequential recursion. Harnessing-level recursion then extracts experience from these edits to update a continual improvement harness, enabling the agent to better adapt to future tasks. Across three professional benchmarks and four model-harness configurations, a single step consistently improves GLM-5 and Gemini 3.5 Flash over their base harnesses, yielding gains of up to 11.5 points on WorkBuddy-Bench, 4.5 points on JobBench, and 11.8 percentage points on LAB, with further gains as the task-level recursion budget increases.", "url": "https://wpnews.pro/news/harness-2-recursive-agent-harnessing-for-an-open-world", "canonical_source": "https://arxiv.org/abs/2610.14224", "published_at": "2026-10-15 04:00:00+00:00", "updated_at": "2026-10-15 05:43:54.080133+00:00", "lang": "en", "topics": ["ai-agents", "artificial-intelligence", "large-language-models", "ai-research"], "entities": ["Harness^2", "GLM-5", "Gemini 3.5 Flash", "WorkBuddy-Bench", "JobBench", "LAB", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/harness-2-recursive-agent-harnessing-for-an-open-world", "markdown": "https://wpnews.pro/news/harness-2-recursive-agent-harnessing-for-an-open-world.md", "text": "https://wpnews.pro/news/harness-2-recursive-agent-harnessing-for-an-open-world.txt", "jsonld": "https://wpnews.pro/news/harness-2-recursive-agent-harnessing-for-an-open-world.jsonld"}}