As terminal-based code agents become more prevalent, the accumulation of agent trajectories has grown rapidly, yet realistic, executable environments remain scarce. However, environments are what agent post-training truly requires: each can be re-queried into many verifiable tasks and provides execution feedback, whereas a trajectory is a single frozen demonstration.
Rather than generating environments from scratch, the researchers observed that the tool-execution history in existing trajectories exposes the structure and contents of the environments in which they ran, making it possible to reconstruct those environments from the trajectories themselves. This insight led to the development of Terminal-Universe, a framework that turns each trajectory into a reusable environment and explores it for synthesizing new tasks and continued interactions.
Specifically, Terminal-Universe replays the file operations recorded in a trajectory to restore each file before the agent modified it, yielding a partial workspace. A completion agent then supplies the missing files and dependencies. On this recovered workspace, the system reconstructs the original intent task and synthesizes entirely new ones.
The framework also scales tasks along two complementary axes: breadth and depth. For breadth, it mines directional dependency relations between related environments and synthesizes cross-workspace queries spanning multiple codebases, as developers routinely do in real-world development. For depth, it extends the initial single-turn query into a multi-round session that captures iterative user feedback and requirement refinement via a user agent.
Applied to public terminal agent trajectories, Terminal-Universe produces 37.3k task-sufficient environments. Supervised fine-tuning of Qwen3.5-27B on this corpus improves single-round performance on Terminal-Bench 2.1 by 11.9 points and multi-round performance on EvoCode-Bench v2 MT@4 by 13.8 points.