agentscope-ai/agentscope

This page introduces an external open-source repository. It is not an HDATF product.

What it does

AgentScope is a framework for building and running LLM agents that relies on the model's own reasoning and tool use instead of rigid prompts. Recent versions add realtime voice agents, the A2A protocol and multi-agent pipelines.

How it helps ATF

Its way of running agents on model reasoning rather than fixed orchestration could be a reference for Harness design. Its A2A support and pipelines are worth comparing with how Harness hands work to executors.

License

Apache-2.0 Permissive, with a patent grant. Commercial use is allowed; keep the notices and state your changes.

More in this category

• Yeachan-Heo/oh-my-claudecode
  A multi-agent orchestration layer for Claude Code with a teams-first design. It provides workflows and a CLI so users can run coordinated agents without first learning Claude Code in depth.
• aiming-lab/AutoResearchClaw
  AutoResearchClaw is a staged pipeline that takes a research idea from chat to a written paper, either autonomously or with a human co-pilot. Its experiment stages route work to domain-specific agents, such as physics, biology and statistics executors.
• Donchitos/Claude-Code-Game-Studios
  A set of Claude Code subagents and slash-command skills for game development, organized like a studio with directors, department leads and specialists. Each agent has defined responsibilities, escalation paths and quality gates, while the user makes the decisions.
• THU-MAIC/OpenMAIC
  OpenMAIC generates interactive multi-agent classroom courses from a prompt. Its agent workbench plans, builds and revises course pages from uploaded materials, with server-backed sessions that can be cancelled, resumed and steered.
• Yeachan-Heo/oh-my-codex
  A workflow layer for OpenAI Codex CLI that adds hooks, agent teams, HUD status displays, prompts and workflows. It is mainly designed and tuned for macOS or Linux.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.