Summary

This repository provides a curated, weekly-updated collection of AI agent research papers published in 2026. It targets AI engineers, researchers, and developers seeking to stay abreast of advancements in agent engineering, memory systems, evaluation methodologies, workflows, and security. The primary benefit is a filtered and categorized overview of the rapidly evolving AI agent landscape, saving users time by pre-selecting relevant, recent publications from arXiv.

How It Works

The project maintains a hand-picked list of research papers focused on the AI agent ecosystem, specifically those published from January 2026 onwards. Papers are sourced from arXiv and categorized into key areas: Multi-Agent Systems, Memory & RAG, Evaluation & Observability, Agent Tooling, and AI Agent Security. The collection is updated weekly to reflect the latest research.

Quick Start & Requirements

This repository is a curated list of research papers and does not involve software installation or execution.

Highlighted Details

• Focuses exclusively on AI agent research papers published in 2026.
• Categorizes papers into five core areas: Multi-Agent, Memory & RAG, Eval & Observability, Agent Tooling, and AI Agent Security.
• Updated weekly from arXiv to ensure recency.
• Includes links to papers via arXiv IDs for direct access.

Maintenance & Community

A Discord community is available for contributions and discussion. The project is maintained by VoltAgent.

Licensing & Compatibility

The repository is licensed under the MIT License. The listed papers are the work of their respective authors and are not endorsed or guaranteed by the curators.

Limitations & Caveats

The curators do not audit, endorse, or guarantee the correctness of the listed research papers. The list is a curated collection, not an active software project.