Awesome Agent Papers

Report Abuse

Basic Information

This repository is a curated, regularly updated collection of research papers and resources about Large Language Model (LLM) agents. It compiles a broad survey and annotated bibliographies across core themes such as agent construction, multi-agent collaboration, agent evolution, tools and tool use, security, benchmarks, datasets, ethics, and real-world applications. The repo includes a taxonomy and an overview figure to help readers orient within the field, and it highlights an accompanying survey paper (arXiv:2503.21460). It is intended as a central reference for researchers, graduate students, and practitioners seeking literature, benchmarks, and design patterns for LLM-based agents. The maintainers invite contributions via pull requests or issues and provide citation information for the survey. The content is organized with a table of contents and topical sections to facilitate discovery and navigation.

Links

Categorization

App Details

Features
The README organizes hundreds of annotated entries under reproducible headings including Agent Collaboration, Agent Construction, Agent Evolution, Applications, Datasets & Benchmarks, Ethics, Security, Tools, and Survey. Each entry lists the paper title, year, and a concise note on contributions or findings, enabling quick scanning of trends. The repo provides curated benchmarks and dataset references, cross-cutting surveys, an overview figure, and a structured taxonomy to relate architectural, evaluation, and safety topics. It calls out prominent frameworks and toolchains in the field and summarizes security threats and defenses specific to agent systems. Contribution instructions, a suggested citation for the survey, and a visible history of updates are included to support community-driven maintenance and reproducibility.
Use Cases
The collection lowers the barrier to entry for studying LLM agents by aggregating seminal and recent work in one place and by offering a taxonomy that clarifies research directions and gaps. Researchers can use the lists to find benchmarks, security analyses, tool-integration studies, and application case studies across domains such as robotics, medicine, finance, and simulation. Educators can draw reading lists and students can locate survey and benchmark papers for projects. Practitioners and system designers can consult summaries of tool frameworks, multi-agent protocols, and safety evaluations to inform system choices. The repository’s contribution workflow and citation guidance help keep the resource current and citable for academic and applied work.

Please fill the required fields*