LiveBench leaderboard for DSH — discovers the release the official site currently publishes, then reads its CSV and category JSON, and renders model rankings in an isolated same-origin iframe opened from a button beside the composer. Search, organization / open-weight / reasoning filters, per-category sorting, task-level detail, up to four models side by side, CSV export; refreshes every 60 s by default (configurable 30–3600) and caches only to the plugin's own .cache/rankings.json.
Install
# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)
dsh plugin --profile web add github:addie-ace/dsh-livebench-rankings
Any plugin you install runs third-party code with your own permissions — it can read your files, use your credentials, and reach the network, and tool approvals don’t sandbox it. GitHub-sourced plugins also run build scripts at install time — pnpm blocks those until you allow them, so an install can stop with ERR_PNPM_GIT_DEP_PREPARE_NOT_ALLOWED or ERR_PNPM_IGNORED_BUILDS; dsh prints the exact key to add under allowBuilds in your profile’s pnpm-workspace.yaml, and the install works on the next run. Allowing a build is a trust decision: only install sources you trust, and pin a commit (github:owner/repo#sha).
README
This plugin publishes its README in Chinese only.
LiveBench Rankings for DeepSeek Harness
独立的 AI 能力排行榜插件。通过 Harness 输入框左下角的「AI 排行」入口打开。
预览
功能
- 从 https://livebench.ai/ 识别当前前端发布的最新榜单版本,再读取官方 CSV 和分类 JSON。
- 默认每 60 秒检查一次,支持手动刷新。不是网站推送:官网/CDN 尚未发布的数据无法提前获取。
- 同步时间和榜单版本分别显示;网络失败时明确提示缓存,不会伪造或填充评分。
- 支持机构、开放权重、推理模型筛选,模型搜索、分项排序、全部变体、微调模型。
- 默认每个模型家族保留综合得分最高的变体,最多对比 4 个模型;可查看任务级评分与导出 CSV。
安装
已构建的发行包没有运行时 npm 依赖。Node.js >= 20.18.1;Windows 使用系统 curl.exe 访问 HTTPS 并校验证书。
在源码工作区运行:
powershell -ExecutionPolicy Bypass -File .\scripts\install.ps1 -HarnessHome 'G:\deepseek harness'
脚本只向 plugins/dsh-livebench-rankings 新增文件,再用官方 dsh plugin --profile web add link:... 注册。
已有同名安装会中止,不覆盖;注册前备份 profile 的 package.json、锁文件和补丁文件。
安装后刷新 Harness 页面;如果客户端入口尚未加载,正常退出并重新打开 Harness。安装脚本不会结束进程或中断对话。
卸载(以本机实际 CLI 为准):
$env:DSH_HOME = 'G:\deepseek harness'
dsh plugin --profile web remove dsh-livebench-rankings
隔离与安全
- 唯一包名、Cordis 行 ID、客户端模块和插槽 ID:
dsh-livebench-rankings。 - 唯一路由前缀:
/dsh-livebench-rankings。不注册全局模型工具、不修改系统提示词、不监听会话内容。 - 原生 dialog 内嵌同源独立 iframe,排行榜 CSS 不进入宿主页面,现有皮肤 CSS 也不会进入榜单。
- 不修改全局 fetch、TLS、代理、路由或 localStorage。不读凭据、账号数据和聊天记录。
- 只从 livebench.ai 下载页面和数据,不执行远程 JavaScript。使用 Acorn 读取静态字面量,Papa Parse 解析 CSV。
- 缓存仅位于插件自身的
.cache/rankings.json。同一实例并发合并,手动刷新至少相隔 5 秒,所有请求有超时和大小限制。 - 卸载时注销路由、清理周期任务、终止进行中的请求、移除对话框和事件监听。
配置
通过 Harness 插件配置编辑器,给本插件独立行设置 refreshSeconds(30–3600,默认 60),可选 proxy(HTTP/SOCKS 代理 URL 或 direct)。不需要修改其他插件。
数据口径
总分是每个能力类别平均分的等权平均,保留两位小数;类别内忽略缺失任务而不是当作零。某个类别完全缺失的模型不进入总榜,与当前官网一致。模型名称、机构、开放权重、推理标记和变体来自官方脚本元数据;历史 Grok-3 特例与官网保持一致。未知模型元数据不猜测。只展示当前官方版本,不将不同版本的分数混为排名变化。
官方前端结构改变时,解析器会报错并保留之前的有效缓存,需要更新适配器。
开发与验证
npm ci
npm run build
npm test
npm run verify:live
npm run dev
预览地址:http://127.0.0.1:4317/dsh-livebench-rankings/ 。生产插件复用 Harness 的服务器,不占用额外端口。
Links
More in this category
zhu1090093659/dsh-web#packages/dsh-task-board★ 7488
Task board for the dsh web GUI: a sidebar multi-column kanban whose cards run in real DSH agent sessions and can also be scheduled with cron expressions, executed host-side even with the browser closed.
zhu1090093659/dsh-web#packages/dsh-web-all★ 7488
Plugin and skin collection for the DSH Web UI: task board, Git graph, right-side panel, remote mobile UI, pet, live token stats, and a skin center.
omdsh-dev/DSH-better-sidebar★ 3555
Full sidebar workbench with file rendering and editing, terminal, Git, and subagents; third-party plugins can register new tabs.
ccch1mneyyy/dsh-TUI★ 2994
Claude Code-style full-screen terminal UI: pixel-whale header, live status line, and streaming thought expansion.
MeteorNOX/DeepSeek-Balance-Whale-Widget★ 2274
A fixed-corner whale widget showing DeepSeek balance, today usage, per-turn cost and random talk lines, with sound effects and a menu.
Devin-AXIS/deepseek-design#deepseek-idesign★ 1072
Visual design studio for websites, app prototypes, posters, cards, reports, and magazines, with templates, direct element editing, selection-aware AI draft handoff, and export.
Community comments
Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.