Abstract
Text-to-SQL models are commonly trained to map questions directly to static queries, whereas real-world database agents operate through stateful, multi-turn interaction with live databases -- inspecting schemas, executing probe queries, diagnosing errors, and revising hypotheses. This creates a critical train-deploy mismatch, as the execution harness that mediates this interaction is introduced only at inference time. To bridge this gap, we propose HarnessSQL, a harness-native post-training framework that preserves the full interaction structure throughout both supervised fine-tuning and reinforcement learning. HarnessSQL builds isolated, executable database environments paired with hidden execution oracles, rolls out teachers directly inside the target SQL harness, and retains only verified trajectories for full-sequence SFT, followed by execution-reward RL. Across Spider 2.0-SQLite, HarnessSQL dramatically boosts the execution accuracy of compact models, raising Qwen3-8B from 15.5% to 45.2% and Qwen3-14B from 22.2% to 54.8%, while transferring effectively to out-of-distribution interactive benchmarks such as BIRD-Interact and LiveSQLBench. Our findings demonstrate that training database agents directly within their execution harness is essential for mastering complex, long-horizon database workflows.
Keywords
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Yang, H., Zhang, J., Xie, J., Gong, S., Han, S., & Guo, Y. (2026). HarnessSQL: Harness-Native Training for SQL Agents in Realistic Database Environments. https://omanscience.com/en/articles/harnesssql-harness-native-training-for-sql-agents-in-realistic-database-environments
MLA 9
Yang, Haolin, et al. "HarnessSQL: Harness-Native Training for SQL Agents in Realistic Database Environments." https://omanscience.com/en/articles/harnesssql-harness-native-training-for-sql-agents-in-realistic-database-environments.
Chicago (author–date)
Yang, Haolin, Jipeng Zhang, Jian Xie, Shuaishuai Gong, Sirui Han, and Yike Guo. 2026. "HarnessSQL: Harness-Native Training for SQL Agents in Realistic Database Environments." https://omanscience.com/en/articles/harnesssql-harness-native-training-for-sql-agents-in-realistic-database-environments.
Harvard
Yang, H., Zhang, J., Xie, J., Gong, S., Han, S. and Guo, Y. (2026) 'HarnessSQL: Harness-Native Training for SQL Agents in Realistic Database Environments', Available at: https://omanscience.com/en/articles/harnesssql-harness-native-training-for-sql-agents-in-realistic-database-environments.
Vancouver
Yang H, Zhang J, Xie J, Gong S, Han S, Guo Y. HarnessSQL: Harness-Native Training for SQL Agents in Realistic Database Environments. https://omanscience.com/en/articles/harnesssql-harness-native-training-for-sql-agents-in-realistic-database-environments
IEEE
H. Yang, J. Zhang, J. Xie, S. Gong, S. Han, and Y. Guo, "HarnessSQL: Harness-Native Training for SQL Agents in Realistic Database Environments," https://omanscience.com/en/articles/harnesssql-harness-native-training-for-sql-agents-in-realistic-database-environments.