# ContinualSkillBench: Can LLM Agents Truly Evolve Their Capabilities?

arXiv: https://arxiv.org/abs/2608.03874 · HTML: https://mlsanity.com/paper/2608.03874/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Placed in Agent skill learning and evolution, one of 16 directions in LLM agents and tool use.
- Builds on 14 of its 49 references that are in the map.
- First posted on arXiv 4 Aug 2026.

## Placed in

- LLM agents and tool use / [Agent skill learning and evolution](https://mlsanity.com/agents/agent-skill-learning-and-evolution/)

## Builds on (most cited first, 14 of 14)

- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023)
- [OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems](https://mlsanity.com/paper/2402.14008/) (2024)
- [GAIA: a benchmark for General AI Assistants](https://mlsanity.com/paper/2311.12983/) (2023)
- [Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces](https://mlsanity.com/paper/2601.11868/) (2026)
- [Solving Challenging Math Word Problems Using GPT-4 Code Interpreter with Code-based Self-Verification](https://mlsanity.com/paper/2308.07921/) (2023)
- [MathCoder: Seamless Code Integration in LLMs for Enhanced Mathematical Reasoning](https://mlsanity.com/paper/2310.03731/) (2023)
- [Claw-Eval: Towards Trustworthy Evaluation of Autonomous Agents](https://mlsanity.com/paper/2604.06132/) (2026)
- [SkillLearnBench: Benchmarking Continual Learning Methods for Agent Skill Generation on Real-World Tasks](https://mlsanity.com/paper/2604.20087/) (2026)
- [Generalization to New Sequential Decision Making Tasks with In-Context Learning](https://mlsanity.com/paper/2312.03801/) (2023)
- [OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning](https://mlsanity.com/paper/2603.08655/) (2026)
- [RealMath: A Continuous Benchmark for Evaluating Language Models on Research-Level Mathematics](https://mlsanity.com/paper/2505.12575/) (2025)
- [PhysicianBench: Evaluating LLM Agents in Real-World EHR Environments](https://mlsanity.com/paper/2605.02240/) (2026)
- [Putnam-AXIOM: A Functional and Static Benchmark for Measuring Higher Level Mathematical Reasoning in LLMs](https://mlsanity.com/paper/2508.08292/) (2025)

## Built on by (most cited first, 8 of 8)

- [S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?](https://mlsanity.com/paper/2608.31100/) (2026)
- [SkillEvoReg: Regularizing Agent Skill Evolution Against Overfitting](https://mlsanity.com/paper/2609.30861/) (2026)
- [ServeLearnBench: How Well Can Agents Self-Improve from Serving Experience?](https://mlsanity.com/paper/2610.07792/) (2026)
- [EmbodiedCLBench: Evaluating Continual Learning for Self-Evolving Embodied Agents](https://mlsanity.com/paper/2610.13021/) (2026)
- [From Existing Methods to New Research Ideas: Two Perspectives on Recursive Self-Improvement](https://mlsanity.com/paper/2610.14625/) (2026)
- [EvoCowork: A Benchmark for Real-World Office Tasks](https://mlsanity.com/paper/2610.14828/) (2026)
- [Socratic Skill Learning: Interleaving Teaching and Practice for Transferable Agent Skills](https://mlsanity.com/paper/2610.17860/) (2026)
- [When Non-Adversarial Noise Hurts Learning: Benchmarking Skill Self-Evolution in Agents](https://mlsanity.com/paper/2610.19910/) (2026)

## Abstract (arXiv)

Modern agent frameworks equip large language models with external skill libraries to solve complex tasks. However, it remains unclear whether these systems can effectively evolve their skills and whether the resulting skills improve task-solving capabilities. To bridge this gap, we introduce ContinualSkillBench, a dynamic evaluation framework for in-context continual skill learning. It covers five domains, each with 100 tasks ordered by difficulty and opportunities for skill transfer. Our experiments show that sequential execution generally improves performance, but the gains vary substantially across models and domains. Moreover, across two models and three domains, retaining context and feedback without skill maintenance yields nearly the same mean performance as maintaining an explicit skill library. Explicit skills nevertheless provide selective benefits for tasks requiring reusable procedures or precise outputs. We further find that the weaker model (GPT-4o) accumulates more skills but reuses them less often, suggesting difficulty consolidating task-specific experience into reusable procedures. These findings show that current in-context skill evolution mechanisms can support continual adaptation, but still struggle to consistently consolidate experience into robust and transferable skills.