HyperSkill: Self-Evolving LLM Agents via Hypergraph-Structured Skill Memory
Organizations: Northwestern University · University of Southern California · University of Illinois at Chicago
Abstract
As agentic tasks grow in complexity, LLM agents increasingly rely on experiential memory to reuse procedural knowledge across tasks. Effective memory design must jointly address what to store, how memory is structured and retrieved, and how memory evolves. Existing systems tackle each only partially: they store trajectories, insights, or workflows as isolated entries, discarding compositional relationships among subtasks and reusable skills; retrieve by flat embedding similarity that ignores relational signals; and maintain memory without leveraging its relational structure. We propose HyperSkill, a hypergraph-based memory framework that jointly improves all three. HyperSkill represents memory as a hypergraph with two node types, subtask steps and reusable skills, where each hyperedge links the subtasks and skills from a single trajectory. Dual-path retrieval queries both subtask and trajectory levels, ranking skills by co-occurrence across retrieved trajectories. Periodic structure-informed maintenance prunes low-utility nodes and merges redundant skills via quality-weighted propagation. Across xBench, GAIA, and WebWalkerQA with GPT-4o and Qwen3-30B-A3B, HyperSkill outperforms ten memory baselines, yielding gains of up to +11.51 on GAIA and +11.18 on WebWalkerQA.
Figures & tables
| Backbone | Method | xBench | GAIA | WebWalkerQA | ||||||
| SR | #Steps | #Calls | SR | #Steps | #Calls | SR | #Steps | #Calls | ||
| Qwen3-30B-A3B | No Memory | 46.00 ↑0.00 | 4.87 ↑0.00 | 5.05 ↑0.00 | 32.12 ↑0.00 | 8.23 ↑0.00 | 4.75 ↑0.00 | 39.41 ↑0.00 | 4.07 ↑0.00 | 3.87 ↑0.00 |
| ReasoningBank | 43.00 ↓3.00 | 4.47 ↓0.40 | 4.72 ↓0.33 | 29.70 ↓2.42 | 7.44 ↓0.79 | 4.42 ↓0.33 | 44.71 ↑5.30 | 3.64 ↓0.43 | 3.72 ↓0.15 | |
| ExpeL | 45.00 ↓1.00 | 5.14 ↑0.27 | 6.06 ↑1.01 | 33.33 ↑1.21 | 7.38 ↓0.85 | 4.15 ↓0.60 | 47.06 ↑7.65 | 4.04 ↓0.03 | 3.94 ↑0.07 | |
| AWM | 44.00 ↓2.00 | 4.86 ↓0.01 | 5.49 ↑0.44 | 32.73 ↑0.61 | 7.07 ↓1.16 | 3.92 ↓0.83 | 41.76 ↑2.35 | 3.84 ↓0.23 | 3.72 ↓0.15 | |
| Generative | 49.00 ↑3.00 | 4.46 ↓0.41 | 4.97 ↓0.08 | 34.55 ↑2.43 | 7.43 ↓0.80 | 3.97 ↓0.78 | 41.18 ↑1.77 | 3.82 ↓0.25 | 3.87 ↑0.00 | |
Appendix figures & tables8 assets
Supplementary material from the paper’s appendix.
Appendix
| Benchmark | GT-judge | Self-judge | Retention |
| xBench | 62.00 | 60.00 | 96.8% |
| GAIA | 44.24 | 38.87 | 87.7% |
| WebWalkerQA | 51.18 | 47.33 | 92.4% |
| What is Stored | How Structured & Retrieved | How Evolved | |||
| Method | Knowledge Form | Compositional? | Structure | Retrieval | Maintenance |
| Voyager ( Wang et al., 2023a ) | Traj. & Tips | ✗ | Vector DB | Semantic | None |
| ExpeL ( Zhao et al., 2024 ) | Traj. & Insights | ✗ | Vector DB | Contrastive | None |
| Generative ( Shang et al., 2025 ) | Traj. & Insights | ✗ | Vector DB | Semantic | None |
| DILU ( Wen et al., 2024 ) | Traj. | ✗ | Vector DB | Semantic | None |
| AWM ( Wang et al., 2025b ) | Workflows | ✗ | In-Prompt | All-in-Prompt | None |
| Memory Scope | Graph Structure | Evolution | ||||
| Method | Mem. Type | MAS | Graph Type | Edge | Retrieval | Maintenance |
| HippoRAG ( Gutiérrez et al., 2024 ) | Factual | ✗ | KG | Pairwise | PPR Activation | None |
| AriGraph ( Anokhin et al., 2025 ) | Factual | ✗ | KG + Episodic | Pairwise | Semantic + Episodic | Graph Update |
| A-Mem ( Xu et al., 2026 ) | Factual | ✗ | Zettelkasten | Pairwise | Semantic + Linking | Self-Organizing |
| Mem0g ( Chhikara et al., 2025 ) | Factual | ✗ | KG + Vector | Pairwise | Hybrid | Deduplication |
| Zep ( Rasmussen et al., 2025 ) | Factual | ✗ | Temporal KG | Pairwise | Temporal + Semantic | Temporal Inval. |
Agent Memory
Agent memory lets AI systems retain, update, and use facts, experiences, and skills across sessions and long-horizon tasks.
See all 190 papers