mlsanityData to 14 Oct 2026

Trace2Tower: Transition-Aware EigenTrace Induction of Multi-Level Skills for LLM Agents

arXiv 2609.05261 ↗First posted 4 Sep 2026cs.AI cs.SEcited by 0

On the map: LLM agents and tool use → Agent skill learning and evolution

Abstract (arXiv)

Large language model agents increasingly rely on execution traces to master complex interactive tasks. However, current paradigms are bottlenecked by shallow trajectory retrieval and flat skill summarization, fundamentally ignoring the temporal dependencies and outcome-conditioned topology of agent behavior. We introduce Trace2Tower, a transition-aware EigenTrace framework that distills raw trajectories into a robust skill hierarchy. Trace2Tower abstracts step-level interactions into canonical events, constructing a unified graph governed by semantic compatibility, transition dynamics, and outcome evidence. Through a novel contrastive spectral decomposition, it isolates stable, success-aligned behavioral modes while rigorously suppressing failure-prone shortcuts. These modes organically populate a dynamic skill tower of action templates, procedural routines, and overarching task strategies, continuously refined via verifier-guided feedback. On ALFWorld, Trace2Tower achieves 87.31% success requiring only 10.35 steps and 0.26 invalid actions; on WebShop, it reaches 50.67% exact success. Across both benchmarks, Trace2Tower significantly outperforms existing baselines in task mastery and context-efficient experience reuse.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Agent skill learning and evolution, one of 16 directions in LLM agents and tool use. Builds on 20 of its 27 references that are in the map. First posted on arXiv 4 Sep 2026.

Builds on

20 of 27 references.

Closest ancestors · 7

Same direction, newest first

  1. 2026SkillDAG: Self-Evolving Typed Skill Graphs for LLM Skill Selection at Scale2 Jun
  2. 2026SkillGraph: Skill-Augmented Reinforcement Learning for Agents via Evolving Skill Graphs12 May
  3. 2026GraSP: Graph-Structured Skill Compositions for LLM Agents20 Apr
  4. 2026Graph-of-Skills: Dependency-Aware Structural Retrieval for Massive Agent Skills7 Apr
  5. 2026Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills26 Mar
  6. 2026SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks13 Feb
  7. 2025Reinforcement Learning for Self-Improving Agent with Skill Library18 Dec

Other references · 13

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  2. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
  3. 2022WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agentscited by 1,251
  4. 2020ALFWorld: Aligning Text and Embodied Environments for Interactive Learningcited by 1,205
  5. 2023ExpeL: LLM Agents Are Experiential Learnerscited by 876
  6. 2024HiAgent: Hierarchical Working Memory Management for Solving Long-Horizon Agent Tasks with Large Language Modelcited by 106
  7. 2025How Memory Management Impacts LLM Agents: An Empirical Study of Experience-Following Behaviorcited by 101
  8. 2025LifelongAgentBench: Evaluating LLM Agents as Lifelong Learnerscited by 78
  9. 2024AgentBank: Towards Generalized LLM Agents via Fine-Tuning on 50000+ Interaction Trajectoriescited by 56
  10. 2025Coarse-to-Fine Grounded Memory for LLM Agent Planningcited by 18
  11. 2026Agent-Diff: Benchmarking LLM Agents on Enterprise API Tasks via Code Execution with State-Diff-Based Evaluationcited by 13
  12. 2025Self-Guided Function Calling in Large Language Models via Stepwise Experience Recallcited by 7
1 more
  1. 2026Why Does the LLM Stop Computing: An Empirical Study of User-Reported Failures in Open-Source LLMscited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026HiSkill: Empowering LLM Agents with Hierarchical Skill Graphs
  2. 2026SkillCommit: Evolving Agent Skills through Behaviorally Validated Scope Expansion
  3. 2025STeCa: Step-level Trajectory Calibration for LLM Agent Learning
  4. 2026SKT: Skill-Use Training at Scale via Verified Synthetic Data Generation
  5. 2026Rep2Skill: Representation-Guided Skill Self-Evolution for LLM Agents
  6. 2026From Traces to Agentic Worlds: Agentic Language World Models for Interactive Environment Simulation
  7. 2026OdysseyArena: Benchmarking Large Language Models For Long-Horizon, Active and Inductive Interactions
  8. 2026SkillAdaptor: Self-Adapting Skills for LLM Agents from Trajectories