Change2Task: From Repository Changes to Executable Coding Agent Tasks and Environments

2026-07-30

AI Summary

This work addresses the limitations in training and evaluating coding agents, which stem from a scarcity of executable, verifiable tasks that reflect real-world software environments. The authors propose a novel paradigm that automatically transforms merged pull requests from code repositories into coding tasks—such as bug fixes and feature additions—that can be executed against modern codebases. By leveraging Patch Reversal, Code Mapping, and Agent Reconstruction techniques, the approach aligns historical code changes with current repository states, enabling end-to-end validation of task lifecycles. Evaluated on 1,130 source changes, the method achieves a 79.6% task construction success rate, recovering 29.2% more tasks than baseline approaches, with 98.0% result consistency and a 10.8% reduction in pipeline overhead.

Abstract

Scaling coding agents requires a continuing supply of executable data for training, benchmarking, and continuous evaluation. Each task must couple a realistic software state with a specification, development tools, and reliable verification. To expand this supply, we present Change2Task, a system grounded in repository history that converts merged pull requests into verified tasks on healthy modern revisions of the same repository. It aligns historical evidence with evolved code, reconstructs task states through Patch Reversal, Code Mapping, or Agent Reconstruction, and validates the lifecycle from a healthy base to a task state and a restored state. By deriving multiple tasks grounded in developer evidence from maintained environments, Change2Task provides executable data for coding agent training and evaluation while reducing repeated environment setup, storage, and task construction effort. We evaluate the system through five common and widely adopted coding agent task families: Bug Fix, Feature Addition, Test Generation, Application Programming Interface Migration, and Security Repair. Starting from 1,130 source changes eligible for construction, Change2Task achieves 79.6% verified task construction success across these task families. On a matched candidate set, it recovers 29.2% more verified tasks than a construction baseline based on pull requests. Historical and reconstructed cases achieve up to 98.0% matched outcome agreement under agent evaluation, while reuse of modern bases reduces measured expenditure across the complete pipeline by 10.8%.

Problem

• coding agents
• executable tasks
• repository changes
• task construction
• software state

Innovation

• Change2Task
• executable coding tasks
• repository history
• task reconstruction
• coding agent evaluation