Activity
Stars and mentions · Last 30 days, Oct 6 – Nov 4
Top repositories
AgnesAI-Labs’s repositories ranked by new stars, as on trending.
Contributes to
Repositories outside AgnesAI-Labs’s account where AgnesAI-Labs opened or reviewed pull requests in the past 365 days.
Who noticed AgnesAI-Labs
Everyone who starred or forked AgnesAI-Labs’s repositories since the archive began in 2011.
Notable accounts
Ordered by the stars their own repositories have received.
Biggest fans
Accounts that starred the most of these repositories.
Breakout weeks
Each repository’s best week of new stars, with the release published around it.
Achievements
Awards earned by the repositories AgnesAI-Labs owns. Select an award to embed its badge.
No awards yet
Repositories earn awards by reaching the top three in a completed trending day or week.
A repository earns an award for finishing first, second or third in a completed trending day or week, ranked by new stars, new forks or pull request authors. Awards follow a repository through renames and transfers and count toward its current owner. Weeks run Monday to Sunday, UTC.
Mentions
Execution is only part of the picture. Our RSI research explores how task feedback can improve future workflows. Agnes Harness: AgnesAI-Labs/agnes-harness Agnes Harness MAS Mode report: github.com/AgnesAI-Labs/Agnes-Harness-MAS-Mode/bl… Build with Agnes.
Prompt: Please help me deploy Agnes Harness : AgnesAI-Labs/agnes-harness AgnesCode: agnes-ai.com
Agnes Harness 的游戏案例里，AI 会打《王者荣耀》，打完还会复盘改策略。大模型负责指挥，Jev 选动作，教练模型读对局记录、修改提示词和执行代码，通过预先设定的测试才启用新版本。官方报告中，对内置机器人打 1v1，胜率从 3% 提到 90%，整个过程没更新模型权重。 AgnesAI-Labs/Agnes-RSI-Framework-Honor-of-Kings
这几天，南京的物理 AI 与世界模型大会上，在一众参展的公司里，Agnes AI 展台前格外热闹。 现场有机械狗和瓦力一起互动，各大屏幕上放着用 AI 打王者荣耀、机械臂抓取等有趣的案例。 这些演示案例背后，用的是他们不久前开源的 Agnes Harness 框架。 GitHub：AgnesAI-Labs/agnes-harness 它把大模型当大脑，决策模型 Jev 当小脑，自己负责记任务进度、调工具，再通过一套叫 MHS 的开放协议，把真实硬件接进来。 在 Agnes Harness 里输入一句话，让机械狗去找好朋友瓦力打个招呼，它就自己走到瓦力跟前挥挥爪子，接着又让它给瓦力跳了个舞。 另外 Agnes Harness 还跟 Agnes Embed TTT 构建了一套双进化智能系统，分为系统和模型进化。 系统进化，现场拿 AI 打王者荣耀作为演示，对局里大模型实时规划策略，每打完一局再复盘，自己调整策略和执行代码，最后胜率 100%。 模型进化交给 Agnes Embed TTT，以机械臂抓取的模拟实验为例，边抓边收集数据拿去训练，实时调整模型的权重参数，越抓越准。 想要了解 AI 是怎么从屏幕走到现有这些物理设备上的朋友，这个开源框架值得研究看看。
别为每个客户复制一套 Agent：Agnes Harness 把业务差异放进插件，让执行、审批和记录留在同一套底座。 AgnesAI-Labs/agnes-harness 这个拆分值得关注：Web、CLI、SDK 共用后端会话，换入口不丢上下文；插件启用时校验完整性与能力哈希，命令默认先审批，执行过程可追溯。业务逻辑和任务方法能沉淀成插件与 Skills，减少下一次交付的重复劳动。 边界也明确：普通后端插件是可信进程内代码，不等于沙箱隔离。目前仍是开发者预览，适合验证架构，别直接当生产承诺。