SkillForge: Self-Distilling Agents for Project-Specific Issue Resolution
Summary
SkillForge is a self-distilling framework that enhances LLM-based software issue resolution by proactively synthesizing and resolving synthetic repository issues to acquire project-specific knowledge, improving performance over baselines.
View Cached Full Text
Cached at: 08/20/26, 04:04 AM
Paper page - SkillForge: Self-Distilling Agents for Project-Specific Issue Resolution
Source: https://huggingface.co/papers/2608.18933
Abstract
SkillForge improves automated software issue resolution by synthesizing repository-specific issues and distilling reusable skills from them before encountering real bugs.
Large language model (LLM) based agents have demonstrated remarkable proficiency in automatedsoftware 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, aself-distillationframework 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 thesesynthetic issues, SkillForge distills reusable project-specific knowledge intoentity-grounded skillsand associates them with relevantrepository entitiesfor 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 downstreamsoftware issue resolution.
View arXiv pageView PDFAdd to collection
Get this paper in your agent:
hf papers read 2608\.18933
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2608.18933 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2608.18933 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2608.18933 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
SkillForge: Co-Evolving Skills and Agents via Dynamic Skill Lifecycles
SkillForge proposes an agentic reinforcement learning method that co-evolves an LLM agent and its skill library through a fitness-driven lifecycle of trial, active, stable, and retired states. It reports up to 7.8% relative improvement over strong baselines on interactive agent benchmarks and releases SkillFurnace, a 5k+ record dataset for studying skill quality and lifecycle management.
Skill Dreaming: Evolving Repository Skills through Pull Request Streams
Skill Dreaming is a framework that extracts executable knowledge from a repository's pull request streams to build and continuously refine a repository-specific skill library for LLM coding agents, enabling co-evolution of PRs, repositories, and skills.
SKILL-DISCO: Distilling and Compiling Agent Traces into Reusable Procedural Skills
Skill-DisCo is a framework that distills reusable procedural skills from successful agent traces and compiles them into callable, executable programs. Experiments on ALFWorld and WebArena show improved success rates and reduced agent turns.
RESOURCE2SKILL: Distilling Executable Agent Skills from Human-Created Multimodal Resources
RESOURCE2SKILL is a framework that distills executable agent skills from multimodal resources like tutorial videos, code repositories, articles, and artifacts into a hierarchical SkillWiki, improving agent performance by 11.9 percentage points over no-skill agents.
SkillGen: Verified Inference-Time Agent Skill Synthesis
This article introduces SkillGen, a multi-agent framework that synthesizes and verifies reusable inference-time skills for LLM agents by contrasting successful and failed trajectories. The method ensures skills are auditable and empirically verified for their net positive impact on agent performance.