mlsanityData to 7 Oct 2026

Towards a Systems Foundation for Agentic Skills: Architecture, Lifecycle, and Security

arXiv 2608.29596 ↗First posted 30 Aug 2026cs.AI cs.LG cs.MAcited by 0

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

Abstract (arXiv)

Autonomous large language model (LLM) agents increasingly face reliability, context consumption, and execution stability bottlenecks when deployed on complex, long-horizon tasks. While monolithic prompt engineering and stateless tool-calling paradigms struggle to scale, the field is rapidly converging toward agentic skills: modular procedural abstractions that externalize execution knowledge into reusable, executable, and portable artifacts. This paper establishes a unified systems foundation and reference architecture for the agentic skills ecosystem. We formalize skills as externalized procedural knowledge bridging high-level cognitive planning with deterministic execution environments, and systematically delineate the architecture across a nine-stage lifecycle: autonomous discovery, authoring and representation formats, memory storage, dynamic retrieval and routing, composition and orchestration, execution and repair, lifelong adaptation, empirical evaluation, and security governance. We further examine marketplace dynamics, public registries, and emerging adversarial threat vectors, alongside runtime verification and defense mechanisms. Finally, we categorize system implementations across software engineering, operating system navigation, embodied robotics, and scientific discovery, while highlighting critical open challenges in continual learning and benchmark realism. This work establishes agentic skills as a foundational paradigm for building scalable, robust, and verifiable autonomous language agents.

Key facts. No citing papers in the map yet (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 137 of its 156 references that are in the map. First posted on arXiv 30 Aug 2026.

Builds on

137 of 156 references.

Closest ancestors · 9

Same direction, newest first

  1. 2026SkillGraph: Self-Evolving Multi-Agent Collaboration with Multimodal Graph Topology19 Apr
  2. 2026The Long-Horizon Task Mirage? Diagnosing Where and Why Agentic Systems Break13 Apr
  3. 2026Agentic Proposing: Enhancing Large Language Model Reasoning via Compositional Skill Synthesis3 Feb
  4. 2026Meta Context Engineering via Agentic Skill Evolution29 Jan
  5. 2026When Single-Agent with Skills Replace Multi-Agent Systems and When They Fail8 Jan
  6. 2025CASCADE: Cumulative Agentic Skill Creation through Autonomous Development and Evolution29 Dec
  7. 2023Cognitive Architectures for Language Agents5 Sep
  8. 2023AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors21 Aug
  9. 2023MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework1 Aug

Other references · 128

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,072
  4. 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,500
  5. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,193
  6. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,155
  7. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,081
  8. 2023AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversationcited by 1,906
  9. 2024SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineeringcited by 1,628
  10. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,302
  11. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
  12. 2023Mind2Web: Towards a Generalist Agent for the Webcited by 1,210
38 more
  1. 2023ChatDev: Communicative Agents for Software Developmentcited by 893
  2. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 751
  3. 2023API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMscited by 574
  4. 2022Visual Programming: Compositional visual reasoning without trainingcited by 507
  5. 2024WebVoyager: Building an End-to-End Web Agent with Large Multimodal Modelscited by 424
  6. 2023Android in the Wild: A Large-Scale Dataset for Android Device Controlcited by 314
  7. 2022ScienceWorld: Is your Agent Smarter than a 5th Grader?cited by 298
  8. 2026SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Taskscited by 289
  9. 2026SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learningcited by 252
  10. 2025SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skillscited by 174
  11. 2023An LLM Compiler for Parallel Function Callingcited by 170
  12. 2026Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skillscited by 161
  13. 2024EASYTOOL: Enhancing LLM-based Agents with Concise Tool Instructioncited by 153
  14. 2023TaskBench: Benchmarking Large Language Models for Task Automationcited by 138
  15. 2026MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agentscited by 129
  16. 2024AnyTool: Self-Reflective, Hierarchical Agents for Large-Scale API Callscited by 121
  17. 2026CoEvoSkills: Self-Evolving Agent Skills via Co-Evolutionary Verificationcited by 117
  18. 2026AutoSkill: Experience-Driven Lifelong Learning via Skill Self-Evolutioncited by 105
  19. 2025Inducing Programmatic Skills for Agentic Taskscited by 103
  20. 2026EvoSkill: Automated Skill Discovery for Multi-Agent Systemscited by 96
  21. 2026Agent Skills in the Wild: An Empirical Study of Security Vulnerabilities at Scalecited by 94
  22. 2025Reinforcement Learning for Self-Improving Agent with Skill Librarycited by 91
  23. 2022Mind's Eye: Grounded Language Model Reasoning through Simulationcited by 83
  24. 2026SKILL0: In-Context Agentic Reinforcement Learning for Skill Internalizationcited by 82
  25. 2017Hierarchical and Interpretable Skill Acquisition in Multi-task Reinforcement Learningcited by 77
  26. 2026SkillClaw: Let Skills Evolve Collectively with Agentic Evolvercited by 76
  27. 2026Skill-SD: Skill-Conditioned Self-Distillation for Multi-turn LLM Agentscited by 75
  28. 2026SWE-Skills-Bench: Do Agent Skills Actually Help in Real-World Software Engineering?cited by 72
  29. 2026XSkill: Continual Learning from Experience and Skills in Multimodal Agentscited by 71
  30. 2026SkillX: Automatically Constructing Skill Knowledge Bases for Agentscited by 69
  31. 2026Agent Skills: A Data-Driven Analysis of Claude Skills for Extending Large Language Model Functionalitycited by 66
  32. 2025Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolutioncited by 65
  33. 2026Organizing, Orchestrating, and Benchmarking Agent Skills at Ecosystem Scalecited by 65
  34. 2026How Well Do Agentic Skills Work in the Wild: Benchmarking LLM Skill Usage in Realistic Settingscited by 62
  35. 2026SkillRouter: Skill Routing for LLM Agents at Scalecited by 57
  36. 2026Memento-Skills: Let Agents Design Agentscited by 56
  37. 2026Graph-of-Skills: Dependency-Aware Structural Retrieval for Massive Agent Skillscited by 53
  38. 2026SkillLearnBench: Benchmarking Continual Learning Methods for Agent Skill Generation on Real-World Taskscited by 45

50 most cited of 128 All 128 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026SoK: Agentic Skills -- Beyond Tool Use in LLM Agents
  2. 2026A Comprehensive Survey on Agent Skills: Taxonomy, Techniques, and Applications
  3. 2026Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward
  4. 2026The Path Ahead for Agentic AI: Challenges and Opportunities
  5. 2025A Survey on Autonomy-Induced Security Risks in Large Model-Based Agents
  6. 2026Agentic Artificial Intelligence (AI): Architectures, Taxonomies, and Evaluation of Large Language Model Agents
  7. 2026Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering
  8. 2026A Systematic Survey of Security Threats and Defenses in LLM-Based AI Agents: A Layered Attack Surface Framework