titleSkillForge: Self-Distilling Agents for Project-Specific Issue Resolution
authorsSilin Chen, Han Li, Xiaodong Gu, Yuling Shi, Haibing Guan
date2026-08-19
venuearXiv (preprint, not peer reviewed)
idarXiv:2608.18933
affiliationShanghai Jiao Tong University
upvotes14 (Hugging Face Daily Papers, 2026-11-01)
linksabs pdf hf
abstractLarge language model (LLM) based agents have demonstrated remarkable proficiency in automated software issue resolution, yet they often struggle to resolve issues in a specific repository because they lack project-specific knowledge. Existing self-evolving approaches acquire such knowledge from repository history or online repair trajectories, but they either depend on available historical issue-resolution signals or incur substantial per-issue test-time exploration cost. In this paper, we propose SkillForge, a self-distillation framework that proactively acquires project-specific knowledge from the repository itself. Instead of waiting for real issues to expose project-specific knowledge gaps, SkillForge synthesizes project-specific issues by re-implementing test-covered core functionalities of the repository. By resolving these synthetic issues, SkillForge distills reusable project-specific knowledge into entity-grounded skills and associates them with relevant repository entities for future issue resolution. Extensive experiments using both open-source and closed-source models show that SkillForge consistently improves issue resolution performance over strong baselines. These results demonstrate that proactively acquiring project-specific knowledge before solving real issues substantially improves downstream software issue resolution.

cite

@misc{chen2026skillforge,
  title={SkillForge: Self-Distilling Agents for Project-Specific Issue Resolution},
  author={Silin Chen and Han Li and Xiaodong Gu and Yuling Shi and Haibing Guan},
  year={2026},
  eprint={2608.18933},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2608.18933}
}