Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills

2026-03-26

AI Summary

Existing automated skill acquisition methods often yield fragile or fragmented skills that struggle to support generalization of large language model (LLM) agents in complex tasks. This work proposes Trace2Skill, a novel framework that, for the first time, emulates the holistic inductive reasoning of human experts. By deploying parallel sub-agents to analyze diverse execution trajectories, the framework performs abductive inference and hierarchical integration to distill a unified, conflict-free, and transferable declarative skill library—without requiring parameter updates or external retrieval. Evaluated across spreadsheet manipulation, visual question answering, and mathematical reasoning tasks, Trace2Skill substantially outperforms strong baselines. Notably, it boosts the performance of Qwen3.5-122B by 57.65 percentage points on WikiTableQuestions, demonstrating robust generalization across model scales and out-of-distribution scenarios.

Technology Category

• Multiagent Systems: Multiagent Learning
• Search and Optimization: Learning to Search
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Equipping Large Language Model (LLM) agents with domain-specific skills is critical for tackling complex tasks. Yet, manual authoring creates a severe scalability bottleneck. Conversely, automated skill generation often yields fragile or fragmented results because it either relies on shallow parametric knowledge or sequentially overfits to non-generalizable trajectory-local lessons. To overcome this, we introduce Trace2Skill, a framework that mirrors how human experts author skills: by holistically analyzing broad execution experience before distilling it into a single, comprehensive guide. Instead of reacting sequentially to individual trajectories, Trace2Skill dispatches a parallel fleet of sub-agents to analyze a diverse pool of executions. It extracts trajectory-specific lessons and hierarchically consolidates them into a unified, conflict-free skill directory via inductive reasoning. Trace2Skill supports both deepening existing human-written skills and creating new ones from scratch. Experiments in challenging domains, such as spreadsheet, VisionQA and math reasoning, show that Trace2Skill significantly improves upon strong baselines, including Anthropic's official xlsx skills. Crucially, this trajectory-grounded evolution does not merely memorize task instances or model-specific quirks: evolved skills transfer across LLM scales and generalize to OOD settings. For example, skills evolved by Qwen3.5-35B on its own trajectories improved a Qwen3.5-122B agent by up to 57.65 absolute percentage points on WikiTableQuestions. Ultimately, our results demonstrate that complex agent experience can be packaged into highly transferable, declarative skills -- requiring no parameter updates, no external retrieval modules, and utilizing open-source models as small as 35B parameters.

Problem

• skill distillation
• trajectory analysis
• transferable skills
• LLM agents
• automated skill generation

Innovation

• skill distillation
• trajectory analysis
• transferable agent skills
• inductive reasoning
• LLM agent