What LLM Agents Can Do and How They Work

In this section, we present a carefully selected set of examples to illustrate the key agent capabilities and their underlying mechanisms (through prompt design) across various applications. These examples are summarized in the next Table, with each presented in a self-contained manner.

An overview of representative examples of LLM agents, their application domains, and demonstrated capabilities.
Example Domain Memory Role-play Planning Tool use Cooperation Reflection
Generative Agents Social simulation
The AI Scientist Scientific discovery
CellAgent Biology
The Virtual Lab Medicine
MetaGPT Software engineering
FinAgent Finance
Voyager Game
Mobile Agent Mobile devices
PaLM-SayCan Embodied AI

If you find this work helpful, please consider citing our paper:

@article{hu2025hands,
  title={Hands-on LLM-based Agents: A Tutorial for General Audiences},
  author={Hu, Shuyue and Ren, Siyue and Chen, Yang and Mu, Chunjiang and Liu, Jinyi and Cui, Zhiyao and Zhang, Yiqun and Li, Hao and Zhou, Dongzhan and Xu, Jia and others},
  journal={Hands-on},
  volume={21},
  pages={6},
  year={2025}
}