‹ back to 2026-09-04

Terminal-Universe: Turning Agent Trajectories into Scalable Terminal Environments

arXiv cs.AI LLM Agents Jie Wu, Zhenru Zhang, Beichen Zhang, Xuwu Wang, Yuhui Su, Mouxiang Chen, Peng Wang, Zhihai Wang, Que Shen, Hao Zhou, An Yang, Fei Huang, Yujiu Yang, Dayiheng Liu 2026-09-03

TL;DR - Terminal-Universe reconstructs reusable, executable coding environments from recorded terminal-agent trajectories, then generates new single- and multi-round tasks from them. The resulting 37.3k environments substantially improved a fine-tuned Qwen3.5-27B on two coding-agent benchmarks.

Replays recorded file operations to recover files as they existed before agent modifications, producing a partial workspace.
Uses a completion agent to supply missing files and dependencies, then reconstructs original tasks and synthesizes new ones.
Expands task breadth with cross-codebase queries and task depth with iterative, multi-round user feedback.
Fine-tuning improved Terminal-Bench 2.1 by 11.9 points and EvoCode-Bench v2 MT@4 by 13.8 points.

view merged work →