mlsanityData to 14 Oct 2026

SkillFM: Generating Skills for LLM Agents via Latent Flow Matching

arXiv 2609.39382 ↗First posted 30 Sep 2026cs.AIcited by 0

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

Abstract (arXiv)

Textual skills provide reusable guidance for large language model agents, but existing approaches often rely on manually curated skill banks or reinforcement learning with indirect and delayed feedback. We introduce SkillFM (Skill Flow Matching), a generative framework that synthesizes task-conditioned textual skills directly without test-time skill retrieval. Our framework combines a codec for encoding and reconstructing textual skills in a continuous latent space with a conditional flow model trained using improved MeanFlow. At inference time, the learned velocity field enables single-step latent sampling, and an LLM-based decoder converts the sampled representation into textual guidance for a frozen downstream agent. We evaluate the framework on embodied tasks, question answering, and web shopping. On ALFWorld and Search-QA, our method achieves the best overall performance among the compared vector-based skill approaches. Our analyses further demonstrate that latent skill generation is an effective alternative to retrieval-based skill augmentation. Our code and training skill libraries are available at https://github.com/lulushang999/SkillFM.

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 41 of its 43 references that are in the map. First posted on arXiv 30 Sep 2026.

Builds on

41 of 43 references.

Closest ancestors · 10

Same direction, newest first

  1. 2026Skill-to-LoRA: From Using Skills to Learning Behaviors for Token-Efficient LLM Agents15 Jun
  2. 2026LatentSkill: From In-Context Textual Skills to In-Weight Latent Skills for LLM Agents4 Jun
  3. 2026More Skills, Worse Agents? Skill Shadowing Degrades Performance When Expanding Skill Libraries21 May
  4. 2026SkillGraph: Skill-Augmented Reinforcement Learning for Agents via Evolving Skill Graphs12 May
  5. 2026SkillOS: Learning Skill Curation for Self-Evolving Agents7 May
  6. 2026SkillClaw: Let Skills Evolve Collectively with Agentic Evolver9 Apr
  7. 2026SKILL0: In-Context Agentic Reinforcement Learning for Skill Internalization2 Apr
  8. 2026Dynamic Dual-Granularity Skill Bank for Agentic RL30 Mar
  9. 2026SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learning9 Feb
  10. 2026Skill-Pro: Learning Reusable Skills from Experience via Non-Parametric PPO for LLM Agents2 Feb

Other references · 31

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  4. 2025Qwen3 Technical Reportcited by 8,820
  5. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  6. 2022Scalable Diffusion Models with Transformerscited by 6,054
  7. 2022Flow Matching for Generative Modelingcited by 5,702
  8. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,754
  9. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,955
  10. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,191
  11. 2020Constructing A Multi-hop QA Dataset for Comprehensive Evaluation of Reasoning Stepscited by 1,282
  12. 2025Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Modelscited by 1,261
19 more
  1. 2022WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agentscited by 1,251
  2. 2020ALFWorld: Aligning Text and Embodied Environments for Interactive Learningcited by 1,205
  3. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,128
  4. 2022Measuring and Narrowing the Compositionality Gap in Language Modelscited by 1,075
  5. 2021MuSiQue: Multihop Questions via Single-hop Question Compositioncited by 916
  6. 2023ExpeL: LLM Agents Are Experiential Learnerscited by 876
  7. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 797
  8. 2025Mean Flows for One-step Generative Modelingcited by 561
  9. 2023A Real-World WebAgent with Planning, Long Context Understanding, and Program Synthesiscited by 353
  10. 2026MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agentscited by 142
  11. 2025Improved Mean Flows: On the Challenges of Fastforward Generative Modelscited by 110
  12. 2025Text-to-LoRA: Instant Transformer Adaptioncited by 58
  13. 2025PonderLM: Pretraining Language Models to Ponder in Continuous Spacecited by 44
  14. 2025LLM Latent Reasoning as Chain of Superpositioncited by 39
  15. 2026SHINE: A Scalable In-Context Hypernetwork for Mapping Context to LoRA in a Single Passcited by 25
  16. 2026Latent-GRPO: Group Relative Policy Optimization for Latent Reasoningcited by 10
  17. 2026Co-Evolving Skill Generation and Policy Optimizationcited by 7
  18. 2026Field-Aware Agent Skill Retrievalcited by 4
  19. 2026Parametric Skillscited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026SkillGym: Internalizing Human Skills into LLMs for Real-World Problem Solving
  2. 2026Skill Is Not Document: Query-Conditioned Compatibility for LLM Agent Skill Routing
  3. 2026SkillReason: Reasoning-Enhanced Agent Skill Retrieval for Implicit User Requests
  4. 2026SkillRet: A Large-Scale Benchmark for Skill Retrieval in LLM Agents
  5. 2025SkillFlow: Scalable and Efficient Agent Skill Retrieval System
  6. 2026SkillCommit: Evolving Agent Skills through Behaviorally Validated Scope Expansion
  7. 2025Automated Skill Discovery for Language Agents through Exploration and Iterative Feedback
  8. 2026SkillAdaptor: Self-Adapting Skills for LLM Agents from Trajectories