Awesome Human Agent Collaboration Interaction Systems

Report Abuse

Basic Information

This repository is an curated, community-maintained collection of literature, code and resources about LLM-based human-agent collaboration and interaction systems. It hosts an organised overview and annotated bibliography centred on a survey paper (LLM-Based Human-Agent Collaboration and Interaction Systems: A Survey), and aggregates recent research papers, open-source repositories, benchmarks, datasets and example applications across multiple domains. The README provides a taxonomy of human feedback, interaction patterns, orchestration strategies and communication structures used in human-agent systems. The project aims to be an exhaustive entry point for researchers and practitioners who want to explore how humans and LLM-powered agents can collaborate, what challenges exist (reliability, complex tasks, safety), and where to find code, benchmarks and domain examples to reproduce, extend or evaluate work in this interdisciplinary field.

Links

Categorization

App Details

Features
The repository organises content by sections that include latest research papers, a detailed taxonomy (Human Feedback, Interaction, Orchestration, Communication), and application-specific collections such as software engineering, embodied AI and robotics, conversational systems, gaming, finance and healthcare. It highlights exemplar projects and benchmarks with dates and code pointers, including MineWorld, SWEET-RL, ConvCodeWorld, FinArena, WebLINX and EmoAgent. The README contains tables summarising feedback types, interaction variants, orchestration strategies and communication modes, plus images and an overview figure. It provides contribution instructions, citation metadata for core survey and related papers, and links to datasets and GitHub projects that serve as benchmarks or codebases for experimentation.
Use Cases
This resource speeds literature review, benchmarking and prototype discovery by centralising recent papers, datasets and open-source implementations related to human-in-the-loop and human-agent collaboration. The taxonomy and tabular summaries help researchers categorise systems by feedback phase, orchestration style, communication structure and interaction type, supporting experimental design and comparison. Practitioners can find domain-specific examples, reproducible benchmarks and dataset pointers to evaluate agent behaviour and safety considerations. Educators and students can use the curated list to build reading lists or course modules. The contribution guidelines and citation entries facilitate community updates and proper attribution when using the survey and compiled resources.

Please fill the required fields*