DeepSeek Harness 3天获12.2万星，GitHub增长最快之一
DeepSeek Harness reached 122K+ GitHub stars in 3 days, one of the fastest to ris…
DeepSeek Harness reached 122K+ GitHub stars in 3 days, one of the fastest to rise in GitHub's history.
DeepSeek Harness 在 3 天内获得了 122K+ 的 GitHub 星标，是 GitHub 历史上增长最快的项目之一。
It makes models, tools, loops, storage, scheduling and even the UI swappable plugins.
它使模型、工具、循环、存储、调度甚至 UI 都成为可替换的插件。
its an orchestration layer for coding agent, runs as a local web app on a configurable port, and exposes every component as a swappable plug-in: shell access, file editing, web search, skills, sessions, model choice, reasoning effort, and permission scope, all editable in a YAML config.
它是编码代理的编排层，作为本地 Web 应用在可配置端口上运行，并将每个组件暴露为可替换的插件：shell 访问、文件编辑、网络搜索、技能、会话、模型选择、推理努力和权限范围，全部可在 YAML 配置中编辑。
sub-agent support is the really special part. You can wire Claude Code or Codex in as plug-ins and let Harness route subtasks to whichever agent suits each step.
子代理支持是特别的部分。你可以将 Claude Code 或 Codex 作为插件接入，让 Harness 将子任务路由到最适合每个步骤的代理。
It's MIT licensed, so you can add other providers or point it at a self-hosted model and never touch DeepSeek's API.
它采用 MIT 许可证，因此你可以添加其他提供商或将其指向自托管模型，而无需接触 DeepSeek 的 API。
However, their release timing for this is strange. They raised the price of cache-hit input by 6-fold to 12-fold in the same week it shipped an agent framework, and cached input is exactly what agent loops consume, since every step replays the same system prompt and accumulated history.
然而，他们的发布时机很奇怪。他们在发布代理框架的同一周将缓存命中输入的价格提高了 6 到 12 倍，而缓存输入正是代理循环所消耗的，因为每一步都会重放相同的系统提示和累积的历史记录。
Whoever set the API pricing either didn't coordinate with the harness group or deliberately priced the new workload higher. My read is the 2nd, because caching subsidies made sense when agents were rare and become the largest unpriced cost once a framework makes them routine.
无论谁设定 API 定价，要么没有与 harness 团队协调，要么故意为新的工作负载定价更高。 我的解读是后者，因为当代理很少时，缓存补贴是有意义的，而一旦框架使它们成为常规，缓存补贴就成为最大的未定价成本。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力