Harness-1: Reinforcement Learning for Search Agents with State-Externalizing Harnesses
2026-06-01 • Artificial Intelligence
Artificial IntelligenceComputation and LanguageInformation Retrieval
AI summaryⓘ
The authors designed a new search agent called Harness-1 that helps decide what information to look for and remember important details during a search. Instead of making the agent handle all the tracking and memory tasks, they created a system (harness) to manage these for it, letting the agent focus on smarter searching choices. This approach improved how well the agent finds useful information across different types of tasks and topics. Their results also suggest that this method helps the agent work well even on new, unseen problems.
search agentreinforcement learningretrievalpolicyworking memorycurated recallmulti-hop QAstate managementtransfer learningenvironment
Authors
Pengcheng Jiang, Zhiyi Shi, Kelly Hong, Xueqiang Xu, Jiashuo Sun, Jimeng Sun, Hammad Bashir, Jiawei Han
Abstract
Search agents are often trained as policies over growing transcripts: the model must decide how to search while also remembering what it has seen, which evidence is useful, which constraints remain open, and which claims have actually been checked. We argue that this formulation puts too much routine state management inside the policy: reinforcement learning is forced to optimize both semantic search decisions and recoverable bookkeeping that the environment can maintain more reliably. We introduce Harness-1, a 20B search agent (retrieval subagent) trained with reinforcement learning inside a stateful search harness. The harness maintains environment-side working memory, including a candidate pool, an importance-tagged curated set, compact evidence links, verification records, compressed and deduplicated observations, and budget-aware context rendering. The policy retains the semantic decisions: what to search, which documents to keep or discard, what to verify, and when to stop. Across eight retrieval benchmarks spanning web, finance, patents, and multi-hop QA, Harness-1 achieves 0.730 average curated recall, outperforming the next strongest open search subagent by +11.4 points and remaining competitive with much larger frontier-model searchers. Its gains are especially strong on held-out transfer benchmarks, suggesting that reinforcement learning over explicit search state can produce retrieval behaviors that generalize beyond the training domains. Our code is available at https://github.com/pat-jj/harness-1.