Search papers, labs, and topics across Lattice.
This paper introduces SkillForge, a self-distillation framework that enhances large language model (LLM) agents by proactively acquiring project-specific knowledge through the synthesis of test-covered core functionalities. Unlike traditional methods that rely on historical issue-resolution signals or costly exploration, SkillForge generates synthetic issues to distill reusable knowledge into entity-grounded skills. Experimental results reveal that SkillForge significantly outperforms existing baselines in software issue resolution, highlighting the importance of preemptively addressing knowledge gaps in specific repositories.
Proactively synthesizing project-specific issues can drastically enhance LLM agents' ability to resolve software problems effectively.
Large 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.
