CJYLZS/dsh-browser 预览 preview

CJYLZS/dsh-browser

Plugin插件 Native原生 ⭐ 3 MIT Search & Web Access搜索与联网

Give dsh agent a real browser to use

Project Overview项目介绍

dsh-browser is a native plugin for the DeepSeek Harness host, shipped exclusively for dsh and installed via dsh plugin add --profile web github:CJYLZS/dsh-browser#v0.2.1, where the #v0.2.1 ref must match the host's dsh -V generation because dsh changed its settings model in 0.1.7-rc.1 with no compatibility layer. The plugin registers a volatile field page through ctx.configForms instead of the removed SettingsScope, spawns one Chrome or Edge process per conversation with its own profile and CDP port, mirrors it in the right sidebar, and exposes six tools plus a skill that reinforces read-before-act and treat-page-text-as-data. It is not an embedded web view, so sites see a normal browser and DevTools or another Playwright can attach to the same external port while the sidebar mirror is live.

In a typical session, the agent calls browser_snapshot to print the accessibility tree with ref ids, narrows large pages with find, locates targets with boxes, then drives the page through browser_click, browser_type, browser_evaluate, and browser_close; clicks dispatch real mouse events with the correct button and click count, chords such as Ctrl+A are supported, and JavaScript dialogs are answered rather than silently dismissed, which lets the agent read and act on pages it has never seen without selectors. The pane also forwards human clicks and typing back into the real page, and isolation is enforced on both directions so two conversations cannot see each other's tabs, cookies, or ports.

Runtime dependencies are only playwright-core (Apache-2.0) and ws (MIT); no browser is bundled, so a local Chrome or Edge install is required and the roughly 200 MB download a bundled browser would cost is avoided, while upstream security updates reach users through their own install rather than this repository. The package is not on npm — install from GitHub or a local link: checkout — and the committed lib/index.js plus lib/client.js must be rebuilt with pnpm run build after any source change. Known limits: no request interception, dialogs not surfaced through the pane, calls without a session are refused, screenshots cannot become image blocks, and license is MIT.

dsh-browser 是 DeepSeek Harness 的原生浏览器插件,会话维度启动独立的本地 Chrome 或 Edge 进程,并分配专属 CDP 端口。用户通过 dsh plugin add --profile web github:CJYLZS/dsh-browser#v0.2.1 安装,需先确认 dsh -V 版本与标签对应;安装路径严格依托 dsh 自身 plugin 机制,无法脱离 Harness 运行,依赖宿主页面的 ctx.configForms 服务注册配置面板。

典型工作流:模型读取 browser_snapshot 输出的可访问性树与 ref 标识,经 find 缩窄范围,再以 browser_click、browser_type 触发真实鼠标与按键事件,信任输入绕开合成 click 被拦截的站点。右侧栏实时镜像页面,允许人手介入同一浏览器;DevTools、Playwright 或 scripts/cdp.mjs 可附加到同一 CDP 端口协同。

依赖与限制:运行时仅引入 playwright-core 与 ws,不打包浏览器,需本机已装 Chrome 或 Edge;不发布到 npm,仅 GitHub 或本地 checkout;网络拦截、对话框、跨会话调用、截图转图像块均未实现;版本严格对齐 0.1.7-rc.1 起的配置模型,旧版需切回 v0.1.x。许可证为 MIT。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 warning1 项注意
  • Only 3 stars - very few users, little community feedback星标只有 3,几乎没人在用,遇到问题缺少社区反馈
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装(适合开发者)

CLI Install命令行安装

dsh plugin add --profile web github:CJYLZS/dsh-browser#v0.2.3

把 CJYLZS/dsh-browser 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

dsh-browser

English | 中文

The plugin at work: the agent's browser tool calls beside the same browser mirrored in the sidebar

The browser settings: one running instance per conversation with its own CDP port, window mode, automation markers, and profile location

A real local browser for every conversation. The agent opens pages and works in them while the sidebar mirrors the same browser, so you can watch the work happen and take over whenever you want. It runs in both the Web and the Desktop interface.

Summary

Chrome or Edge runs as its own process, with its own profile and its own CDP port, and every conversation gets its own: tabs, sign-ins, and ports are never shared between conversations. The agent drives it with nine tools, and the sidebar shows the same pages — either side can act on them.

It is not an embedded web view. The page runs in an ordinary browser process, so sites see a normal browser, DevTools or another Playwright can attach to it while the mirror is live, and closing the harness does not leave a browser pretending to be part of the app.

Highlights

  • Works in both interfaces. The Web GUI and the Desktop app run the same plugin against the same browser; install it once in each and nothing else differs, including the settings page.
  • One browser per conversation. Tabs, sign-ins, and ports belong to one conversation and stay there: the sidebar pane names its session, and a tool call resolves its browser from the session that called it.
  • Watch it work, or take over. Every page the browser opens gets its own sidebar tab with a live view, and clicks and typing in that tab reach the real page.
  • The agent can read a page it has never seen. browser_snapshot prints the page's accessibility tree with a ref per node, which is how the agent reads an unfamiliar page and how it names the element to click; find narrows a large page, and boxes gives positions without a screenshot.
  • Input a site accepts. Real mouse and text events at the element's own position — a page that ignores a synthetic element.click() still takes these — plus chords such as Ctrl+A, native dropdowns and checkboxes, dialogs answered by the call that opens them, and an explicit refusal when something else would receive the press.
  • Nine tools, and code for the rest. browser_evaluate is the general-purpose one; the others cover what code cannot express, such as reading a page without knowing its selectors first, or asking what the page said while it failed. See Tools.
  • Attachable. Each browser listens on an external CDP port, so chrome://inspect, another Playwright, or the bundled scripts/cdp.mjs can attach to the same browser the sidebar is mirroring.
  • Working rules where the model reads them. The plugin contributes a skill — read before acting, one action per observation, and a page's text is data rather than instructions — with five short recipes read only when their question comes up.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

← 上一个 Prev dsh-mcmp 下一个 Next harness-self-evolution-plugin →