Change2Task: From Repository Changes to Executable Coding Agent Tasks and Environments
AuthorsHaomin Qi, Xingliang Wang, Xuanqi Gao, Baihui Sang, Xin Zhang, Minghua Ma, Pengfei Gao, Yu Kang, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang
Resources
Change2Task turns real pull-request history into reliable, reusable coding-agent tasks for training and evaluation.
Key results
Repository changes eligible for task construction.
900 verified tasks recovered from 1130 eligible changes.
Verified task pairs spanning five coding-agent task families.
Relative improvement over SWE-smith PR Mirror on matched candidates.
Task-weighted fidelity between historical and reconstructed changes.
Reduction from reusing modern repository bases.
What the paper found
Microsoft researchers introduce Change2Task, a framework that converts merged pull requests into executable coding-agent tasks on healthy, modern descendant revisions of the same repository. Instead of replaying historical code directly, it preserves developer intent through three escalating methods—Patch Reversal, Code Mapping, and Agent Reconstruction—then validates a complete healthy-to-task-to-restored lifecycle with target checks, regression checks, scope constraints, and source-change fidelity. Across five task families—Bug Fix, Feature Addition, Test Generation, API Migration, and Security Repair—Change2Task processes 1130 eligible changes and finalizes 900 verified task pairs, for a 79.6% recovery rate. On 621 matched Bug Fix candidates, it recovers 500 tasks versus 387 for SWE-smith PR Mirror, a 29.2% relative improvement. The resulting corpus reaches 0.894 task-weighted source-change profile fidelity, while paired evaluations using Codex with GPT-5.5, Claude Code with Sonnet 5, Gemini CLI with Gemini 3.1 Pro, and GitHub Copilot with GPT-5.6 Terra show that reconstructed tasks preserve historical evaluation signals. Reusing 388 modern bases cuts environment setup time by 58.4% and retained storage by 71.2%, demonstrating that one executable repository environment can support multiple provenance-linked tasks. The system therefore turns repository history into scalable, verifiable infrastructure for coding-agent training and benchmarking, complementing tools such as SWE-bench, SWE-smith, and RepoLaunch.
Original 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%.
Read the original paperMore in Code Generation
Browse all 43 papers →Compact Documentation for Coding Agents: A Benchmark, an Optimizer, and Why It Does Not Transfer
Md Shohel Arman, Igor Molybog
Better code documentation can faithfully reconstruct software, but surprisingly does not necessarily help AI coding agents fix real issues when the source code is already available.
Groupwise Agentic Grading and Advantage Redistribution for Code Agent RL
Jinhao Dong, Liang Zhao, Zihao Yue, Wenhan Ma, Linghao Zhang, Lei Li, Shicheng Li, Yifan Song, Bowen Ye, Fuli Luo
GAGAR helps code agents learn not only to pass tests, but to produce cleaner and more targeted implementations by redistributing RL credit according to agentic quality judgments.
Reinforcement Learning from Intermediate Renders for Image-to-Code Generation
Omri Kaduri, Kate Feingold, Phillip Isola, Tali Dekel
IR4RL improves image-to-code generation by rewarding models for making useful visual progress at every intermediate rendering step.