mlsanityData to 7 Oct 2026

CASCADE: Cumulative Agentic Skill Creation through Autonomous Development and Evolution

arXiv 2512.23880 ↗First posted 29 Dec 2025cs.AI cond-mat.mtrl-scicited by 33

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

Abstract (arXiv)

Large language model (LLM) agents currently depend on predefined tools or early-stage tool generation, limiting their adaptability and scalability to complex scientific tasks. We introduce CASCADE, a self-evolving agentic framework representing an early instantiation of the transition from "LLM + tool use" to "LLM + skill acquisition". CASCADE enables agents to master complex external tools and codify knowledge through two meta-skills: continuous learning via web search, code extraction, and memory utilization; self-reflection via introspection, knowledge graph exploration, and others. We evaluate CASCADE on SciSkillBench, a benchmark of 116 materials science and chemistry research tasks. CASCADE achieves a 93.3% success rate using GPT-5, compared to 35.4% without evolution mechanisms. We further demonstrate real-world applications in computational analysis, autonomous laboratory experiments, and selective reproduction of published papers. Along with human-agent collaboration and memory consolidation, CASCADE accumulates executable skills that can be shared across agents and scientists, moving toward scalable AI-assisted scientific research.

Key facts. Cited by 33 papers in the map (data through 7 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 41 of its 111 references that are in the map. First posted on arXiv 29 Dec 2025.

Builds on

41 of 111 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025Adaptation of Agentic AI: A Survey of Post-Training, Memory, and Skills18 Dec
  2. 2025Alita: Generalist Agent Enabling Scalable Agentic Reasoning with Minimal Predefinition and Maximal Self-Evolution26 May

Other references · 39

  1. 2025Qwen3 Technical Reportcited by 8,195
  2. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  3. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,072
  4. 2022Survey of Hallucination in Natural Language Generationcited by 2,056
  5. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,880
  6. 2023Teaching Large Language Models to Self-Debugcited by 963
  7. 2024OpenHands: An Open Platform for AI Software Developers as Generalist Agentscited by 956
  8. 2025Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memorycited by 722
  9. 2023ChemCrow: Augmenting large-language models with chemistry toolscited by 609
  10. 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 467
  11. 2025Agentic Context Engineering: Evolving Contexts for Self-Improving Language Modelscited by 349
  12. 2023Large Language Models as Tool Makerscited by 293
27 more
  1. 2024ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discoverycited by 208
  2. 2025SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skillscited by 174
  3. 2023CRAFT: Customizing LLMs by Creating and Retrieving from Specialized Toolsetscited by 122
  4. 2024Language agents achieve superhuman synthesis of scientific knowledgecited by 121
  5. 2025Kosmos: An AI Scientist for Autonomous Discoverycited by 78
  6. 2025Robin: A multi-agent system for automating scientific discoverycited by 57
  7. 2024Open Materials 2024 (OMat24) Inorganic Materials Dataset and Modelscited by 50
  8. 2025AutoAgent: A Fully-Automated and Zero-Code Framework for LLM Agentscited by 49
  9. 2025LLM Agents Making Agent Toolscited by 45
  10. 2025SciMaster: Towards General-Purpose Scientific AI Agents, Part I. X-Master as Foundation: Can We Lead on Humanity's Last Exam?cited by 45
  11. 2025El Agente: An Autonomous Agent for Quantum Chemistrycited by 43
  12. 2025SciToolAgent: A Knowledge Graph-Driven Scientific Agent for Multi-Tool Integrationcited by 40
  13. 2024CACTUS: Chemistry Agent Connecting Tool-Usage to Sciencecited by 32
  14. 2025ToolUniverse: An open platform for democratizing AI scientistscited by 22
  15. 2025The Denario project: Deep knowledge AI agents for scientific discoverycited by 21
  16. 2025Paper2Agent: Reimagining Research Papers As Interactive and Reliable AI Agentscited by 19
  17. 2025Context Engineering 2.0: The Context of Context Engineeringcited by 17
  18. 2025Accelerating Scientific Discovery with Autonomous Goal-evolving Agentscited by 16
  19. 2025Automating MD simulations for Proteins using Large language Models: NAMD-Agentcited by 12
  20. 2025MatTools: Benchmarking Large Language Models for Materials Science Toolscited by 11
  21. 2024AlabOS: A Python-based Reconfigurable Workflow Management Framework for Autonomous Laboratoriescited by 10
  22. 2025Toward Greater Autonomy in Materials Discovery Agents: Unifying Planning, Physics, and Scientistscited by 10
  23. 2025Advanced Tool Learning and Selection System (ATLASS): A Closed-Loop Framework Using LLMcited by 8
  24. 2025OBELiX: A Curated Dataset of Crystal Structures and Experimentally Measured Ionic Conductivities for Lithium Solid-State Electrolytescited by 5
  25. 2025Operationalizing Serendipity: Multi-Agent AI Workflows for Enhanced Materials Characterization with Theory-in-the-Loopcited by 4
  26. 2025PARC: An Autonomous Self-Reflective Coding Agent for Robust Execution of Long-Horizon Taskscited by 2
  27. 2015User Applications Driven by the Community Contribution Framework MPContribs in the Materials Projectcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 33.

33 papers.

Sort:

By area:

  1. 2026SkillsVote: Lifecycle Governance of Agent Skills from Collection, Recommendation to Evolutioncited by 18
  2. 2026A Comprehensive Survey on Agent Skills: Taxonomy, Techniques, and Applicationscited by 17
  3. 2026SkillGenBench: Benchmarking Skill Generation Pipelines for LLM Agentscited by 17
  4. 2026Experience Compression Spectrum: Unifying Memory, Skills, and Rules in LLM Agentscited by 16
  5. 2026Co-Evolving LLM Decision and Skill Bank Agents for Long-Horizon Taskscited by 13
  6. 2026Agentic World Modeling: Foundations, Capabilities, Laws, and Beyondcited by 10
  7. 2026Library Drift: Diagnosing and Fixing a Silent Failure Mode in Self-Evolving LLM Skill Librariescited by 10
  8. 2026OpenSkill: Open-World Self-Evolution for LLM Agentscited by 9
  9. 2026Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Applicationcited by 8
  10. 2026AutoScientists: Self-Organizing Agent Teams for Long-Running Scientific Experimentationcited by 7
  11. 2026Skill Coverage: A Test Adequacy Metric for Agent Skillscited by 7
  12. 2026Dynamic Agent Skills: A Lifecycle Survey and Taxonomy of Evolving Skill Librariescited by 6
21 more
  1. 2026MADE: Benchmark Environments for Closed-Loop Materials Discoverycited by 5
  2. 2026TritonDFT: Automating DFT with a Multi-Agent Frameworkcited by 5
  3. 2026Harnessing AtomisticSkills for Agentic Atomistic Researchcited by 5
  4. 2026Co-Evolving Skill Generation and Policy Optimizationcited by 5
  5. 2026The Last AI Built by Humans: Toward Genuine Recursive Self-Improvementcited by 5
  6. 2026El Agente Gr'afico: A Semantic Execution Runtime for Scientific Agentscited by 4
  7. 2026SwarmWorld: Stigmergic technological evolution in societies of language-model agentscited by 3
  8. 2026Rooted Absorbed Prefix Trajectory Balance with Submodular Replay for GFlowNet Trainingcited by 2
  9. 2026Mimosa Framework: Toward Evolving Multi-Agent Systems for Scientific Researchcited by 1
  10. 2026Towards Knowledgeable Deep Research: Framework and Benchmarkcited by 1
  11. 2026The Blind Curator: How a Biased Judge Silently Disables Skill Retirement in Self-Evolving Agentscited by 1
  12. 2026Hypothesis-Driven Autonomous Materials Synthesis with Multimodal LLM Agentscited by 1
  13. 2026SearchSkill: Teaching LLMs to Use Search Tools with Evolving Skill Bankscited by 0
  14. 2026DRIVE: Modeling Skills at the Reasoning and Interaction Levels for Web Agents under Continual Learningcited by 0
  15. 2026Harnessing X-ray Absorption Spectroscopy Data through Multimodal Mining of Battery Literaturecited by 0
  16. 2026Quo Vadis, World Modeling?cited by 0
  17. 2026Harnessing agent memory to build lifelong AI partners for materials scientistscited by 0
  18. 2026Towards a Systems Foundation for Agentic Skills: Architecture, Lifecycle, and Securitycited by 0
  19. 2026Agentic programs: an emerging form of scientific software in computational materials sciencecited by 0
  20. 2026Grounded Skill Synthesis from Code at Scale for Agentic Intelligencecited by 0
  21. 2026VideoResearcher: Self-Improving Tool Design for Long-Video Understandingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Agent0: Unleashing Self-Evolving Agents from Zero Data via Tool-Integrated Reasoning
  2. 2026SKILLFOUNDRY: Building Self-Evolving Agent Skill Libraries from Heterogeneous Scientific Resources
  3. 2026MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation
  4. 2026Towards AGI A Pragmatic Approach Towards Self Evolving Agent
  5. 2026ContinualSkillBench: Can LLM Agents Truly Evolve Their Capabilities?
  6. 2026MetaSkill-Evolve: Recursive Self-Improvement of LLM Agents via Two-Timescale Meta-Skill Evolution
  7. 2026Skill-MAS: Evolving Meta-Skill for Automatic Multi-Agent Systems
  8. 2026SkillClaw: Let Skills Evolve Collectively with Agentic Evolver