StateFork and Waypoint improve AI agent exploration in terminal environments
StateFork: Branchable Infrastructure for Agent Exploration
Operating Systems
Summary
AI agents that work on computers try different action paths to solve tasks, but changing files and processes means their environment changes too. The authors introduce StateFork, a system that lets AI safely explore different paths by saving and restoring the computer state efficiently. They also build Waypoint, a tool that checkpoints terminal sessions with file and process states, making it faster and more accurate for AI agents to try out actions. This combination lets AI agents explore more effectively and finish tasks more reliably.
What this means in practice
- •For software automation teams: Improve automated scripts by enabling them to explore multiple action paths and restore system state without restarting from scratch.
- •For devops engineers: Enhance testing and debugging by efficiently branching and restoring terminal environments to reproduce complex scenarios quickly.
Authors
Jiakai Xu, Tianle Zhou, Georgios Liargkovas, Danielle Gillai, Ruizhe Fu, Patrick Shen, Eugene Wu, Kostis Kaffes
Abstract
AI agents improve task success by exploring multiple trajectories, but for computer-use agents each trajectory modifies external environment state. Branching from an intermediate point is correct only when restoration is observation-equivalent - future actions produce the same observations - and practical only when creating, restoring, and discarding branch states is physically efficient. We study this problem for terminal-using agents, where tasks modify files, shell context, running processes, and local services. We introduce StateFork, a logical control plane that separates exploration policies from physical state materialization, exposing sessions, commands, snapshots, restores, and cleanup over multiple execution substrates. We also build Waypoint, a checkpoint/restore substrate for terminal execution sessions that combines filesystem layering, process checkpointing, and a persistent terminal-compatible command session. Together, StateFork and Waypoint improve terminal-agent exploration by combining sample-efficient search with efficient restoration of the right execution state. On Terminal-Bench, branch-based exploration through StateFork and Waypoint improves task completion over pass@20 at the same visited-node budget, and using Waypoint achieves 26% higher task accuracy than other execution substrates while completing exploration up to 70% faster. These results show that observation-equivalent, physically efficient execution sessions are a key systems abstraction for exploratory AI agents.