mlsanityData to 14 Oct 2026

ASPIRE: Agentic /Skills Discovery for Robotics

arXiv 2607.00272 ↗First posted 30 Jun 2026cs.RO cs.AI cs.MAcited by 53

On the map: LLM agents and tool use → Multimodal and embodied agents

Abstract (arXiv)

Traditional robot programming is challenging: it requires orchestrating multimodal perception, managing physical contact dynamics, and handling diverse configurations and execution failures. We introduce ASPIRE (Agentic Skill Programming through Iterative Robot Exploration), a continual learning system that autonomously writes and refines robot control programs in a code-as-policy paradigm while compounding experience into a reusable skill library. ASPIRE discovers skills that persist across tasks, simulation and real-world settings, and embodiments. It operates in an open-ended loop with three components: (1) a closed-loop robot execution engine that exposes fine-grained multimodal traces, enabling autonomous failure diagnosis, repair synthesis, and validation; (2) a continually expanding skill library that distills validated fixes into reusable, transferable knowledge; and (3) evolutionary search that generates diverse task sequences and control programs to explore beyond single-trajectory refinement. ASPIRE surpasses prior methods by up to 77% on LIBERO-Pro manipulation under perturbation, 72% on Robosuite bimanual handover, and 32% on BEHAVIOR-1K long-horizon household tasks. Its accumulated library also enables zero-shot generalization to unseen long-horizon tasks: on LIBERO-Pro Long, ASPIRE achieves 31% success versus 4% for prior methods despite their use of test-time reasoning and retries. Finally, simulation-discovered skills provide initial evidence of sim-to-real transfer, substantially reducing real-robot programming effort across different embodiments and robot APIs.

Key facts. Cited by 53 papers in the map (data through 14 Oct 2026). Placed in Multimodal and embodied agents, one of 16 directions in LLM agents and tool use. Builds on 39 of its 48 references that are in the map. First posted on arXiv 30 Jun 2026.

Builds on

39 of 48 references.

Closest ancestors · 2

Same direction, newest first

  1. 2026CaP-X: A Framework for Benchmarking and Improving Coding Agents for Robot Manipulation23 Mar
  2. 2026RoboClaw: An Agentic Framework for Scalable Long-Horizon Robotic Tasks12 Mar

Other references · 37

  1. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,373
  2. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  3. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  4. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
  5. 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,348
  6. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,191
  7. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 1,761
  8. 2024SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineeringcited by 1,737
  9. 2024Octo: An Open-Source Generalist Robot Policycited by 1,484
  10. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,410
  11. 2025GR00T N1: An Open Foundation Model for Generalist Humanoid Robotscited by 1,225
  12. 2022Code as Policies: Language Model Programs for Embodied Controlcited by 1,210
25 more
  1. 2024OpenHands: An Open Platform for AI Software Developers as Generalist Agentscited by 1,002
  2. 2023Teaching Large Language Models to Self-Debugcited by 988
  3. 2025AlphaEvolve: A coding agent for scientific and algorithmic discoverycited by 867
  4. 2023VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Modelscited by 750
  5. 2022ProgPrompt: Generating Situated Robot Task Plans using Large Language Modelscited by 611
  6. 2023Eureka: Human-Level Reward Design via Coding Large Language Modelscited by 571
  7. 2024Executable Code Actions Elicit Better LLM Agentscited by 564
  8. 2020robosuite: A Modular Simulation Framework and Benchmark for Robot Learningcited by 511
  9. 2023Language to Rewards for Robotic Skill Synthesiscited by 312
  10. 2023RoboGen: Towards Unleashing Infinite Data for Automated Robot Learning via Generative Simulationcited by 218
  11. 2023Guiding Pretraining in Reinforcement Learning with Large Language Modelscited by 187
  12. 2025LIBERO-PRO: Towards Robust and Fair Evaluation of Vision-Language-Action Models Beyond Memorizationcited by 179
  13. 2023Text2Reward: Reward Shaping with Language Models for Reinforcement Learningcited by 160
  14. 2024BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulationcited by 148
  15. 2024DrEureka: Language Model Guided Sim-To-Real Transfercited by 83
  16. 2025MuJoCo Playgroundcited by 78
  17. 2024RoboCodeX: Multimodal Code Generation for Robotic Behavior Synthesiscited by 58
  18. 2024Lifelong Robot Library Learning: Bootstrapping Composable and Generalizable Skills for Embodied Control with Language Modelscited by 27
  19. 2026K-Search: LLM Kernel Generation via Co-Evolving Intrinsic World Modelcited by 24
  20. 2026AtomVLA: Scalable Post-Training for Robotic Manipulation via Predictive Latent World Modelscited by 23
  21. 2024Eurekaverse: Environment Curriculum Generation via Large Language Modelscited by 16
  22. 2025VistaWise: Building Cost-Effective Agent with Cross-Modal Knowledge Graph for Minecraftcited by 13
  23. 2025EvoEngineer: Mastering Automated CUDA Kernel Code Evolution with Large Language Modelscited by 13
  24. 2026Uni-Skill: Building Self-Evolving Skill Repository for Generalizable Robotic Manipulationcited by 9
  25. 2026Code Evolution for Control: Synthesizing Policies via LLM-Driven Evolutionary Searchcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 53.

53 papers.

Sort:

By area:

  1. 2026Harness VLA: Steering Frozen VLAs into Reliable Manipulation Primitives via Memory-Guided Agentscited by 47
  2. 2026Code as Worlds: Agentic Discovery of Executable World Representations for Physical Reasoningcited by 14
  3. 2026The Last AI Built by Humans: Toward Genuine Recursive Self-Improvementcited by 11
  4. 2026Towards the Harness of Embodied Agentscited by 9
  5. 2026Agent as Policy for Robotic Manipulationcited by 8
  6. 2026PhyAgentOS: A Self-Evolving Operating System for Embodied Agents with Decoupled Cognitive Planning and Physical Executioncited by 6
  7. 2026Don't Drop the BATON: Long-Horizon Robot Manipulation via Agentic Subtask Exploration and Transition-aware Memorycited by 5
  8. 2026HarnessPAI: An Evolving Harness for Physical AIcited by 5
  9. 2026EmbodiedSWE: Coding Agents for Long Horizon Dexterous Roboticscited by 4
  10. 2026Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success Rate but 65% Fewer Tokenscited by 4
  11. 2026HarnessVLN: Unifying Training-Free Embodied Navigation through an Agent Harnesscited by 3
  12. 2026Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulationcited by 3
38 more
  1. 2026RACaP: Agentic Reasoning, Acting, and Coding as Policies for Evolvable Robot Learningcited by 3
  2. 2026What Stops Recursive Self-Improvement in Robotics? Lessons from 123 Rounds of Agentic Skill Discoverycited by 3
  3. 2026Learning and Transferring Closed-Loop Robot Softwarecited by 2
  4. 2026Coding Agents for Generalized Task and Motion Planning Problemscited by 2
  5. 2026RAPID: Robot Agentic Programming from Demonstrationscited by 2
  6. 2026Scale and Selection: What Makes Automatic Harness Evolution Work for Visual-Interface Robot Agentscited by 2
  7. 2026EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolutioncited by 2
  8. 2026RoboRSI: Stable, efficient, and reusable robot self-evolution in complex real-world environmentscited by 2
  9. 2026Weights or Skills? A Survey of Robot-Learning Techniques: from Action-Predicting Weights to Robots that Write their Own Skillscited by 1
  10. 2026Teach and Grow: An Agent-Centered Architecture for General Robot Learningcited by 1
  11. 2026SUN: Agentic Robot Policy Learning with Persistent Task Programscited by 1
  12. 2026StageGuard: Learning Stage Transitions for Long-Horizon Robot Tasks via Agentic Distillationcited by 1
  13. 2026Generalizing Manipulation Skills with a Local Coding Agentcited by 1
  14. 2026AdaHVLA: Adaptive Harnesses for Long-Horizon Vision-Language-Action Executioncited by 1
  15. 2026Find Something You Can't Do: Agentic Real-World Reinforcement Learning for Self-Improving VLA Modelscited by 1
  16. 2026RoboFoundry: System-as-Policy Evolution for Self-Learning Embodied Agentscited by 1
  17. 2026EMPIRIC: Experiment-Driven Learning of Residual World Models for Robot Planningcited by 1
  18. 2026Self-Evolving Coding Agents: From Digital Programs to Physical-World Intelligencecited by 1
  19. 2026In-Context Learning for Robots: Methods and Applicationscited by 1
  20. 2026Real2Gym: Building Gyms from Videos, Bringing Skills to Robotscited by 1
  21. 2026EMHO: EMbodied Agent Harness Optimization via Experience Tracescited by 1
  22. 2026RobotWorld: Benchmarking Multimodal Agents for Robot Use Across Diverse Tasks and Embodimentscited by 1
  23. 2026Agentic RSR: Real-to-Sim-to-Real through Scene Reconstruction and Execution-Grounded Robot Policiescited by 1
  24. 2026Embodied Turing Machines: Stateful Code for Robot Recursive Self-Improvementcited by 1
  25. 2026RoboHarness: A Simple Harness Outperforms VLA and World Actions Modelscited by 1
  26. 2026CAPEX: Efficiently Distilling Foundation Model Behavior into Deployable Robot Policies through Experience-Adaptive Reasoningcited by 0
  27. 2026SkillWeaver: Agentic Exploration over Neural Interaction Skills for Scalable Robot Data Generationcited by 0
  28. 2026Encore: Few-Shot Agentic Discovery of Manipulation Strategiescited by 0
  29. 2026Explore, Execute, Evolve: A Skill Acquisition and Reuse Loop for Embodied Agentscited by 0
  30. 2026MotorMind: Scaffolding General Vision Language Models for Zero-Shot Robot Manipulationcited by 0
  31. 2026Skill-Space Shooting for Autonomous Robot Policy Improvementcited by 0
  32. 2026DynaHarness: A Dynamic Physical Harness for Self-Evolving Robot Agentscited by 0
  33. 2026Iterative Policy Refinement through Semantic Rollout Analysiscited by 0
  34. 2026InterEvolve: Test-Time Evolution of Reward Programs for Humanoid Loco-Manipulationcited by 0
  35. 2026BiGym 2.0: Benchmarking Learned and Agent-Developed Policies for Humanoid Household Manipulationcited by 0
  36. 2026PhysEvo: Astra Can Act, Let Itcited by 0
  37. 2026RoboAware: Learning to Coordinate Embodied Skills from Counterfactual Outcomescited by 0
  38. 2026Fly-by-Code: Embodied Coding Agents for Aerial Manipulation with Active Visual and Physical Feedbackcited by 0

50 most cited of 53 All 53 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Reconstruct, Practice, Go Real: Guided Self-Improvement for Embodied Agents
  2. 2026Embodied Skill Bench: Benchmarking How Well Agents with Embodied Skills Perform on Physical Tasks
  3. 2026Skill2Real: Agentic Skill Learning for Zero-Shot Sim-to-Real Robot Manipulation
  4. 2026ENPIRE: Agentic Robot Policy Self-Improvement in the Real World
  5. 2026Playful Agentic Robot Learning
  6. 2024Agentic Skill Discovery
  7. 2026SkillFlow:Benchmarking Lifelong Skill Discovery and Evolution for Autonomous Agents
  8. 2026ASENA: Self-evolving Agents for Embodied Navigation