ben7am1n/dsh-browser
基于 Playwright 的浏览器自动化,专为 DeepSeek Harness 设计
Project Overview项目介绍
This is a browser automation plugin built specifically for DeepSeek Harness, powered by the Playwright browser automation library. It extends DSH with a full suite of browser_* prefixed tools that let the DSH agent interact directly with live web pages. You can install the plugin directly from the DSH command line by running dsh plugin --profile <your-profile-name> add dsh-browser, and no manual build steps are required for most default DSH installations. The browser persists across tool calls, so you only open it once and perform multiple actions before closing it.
The plugin is designed for DSH users that need their agent to complete web-based tasks that require real browser interaction, rather than just simple HTTP fetching of static content. A typical workflow starts with calling browser_open to spin up a browser instance and navigate to an initial target URL. After that, the agent can call a sequence of tools to click elements, type into inputs, select options, wait for slow or lazy-loaded content, and pull text or HTML content from the page for processing. When done, the agent calls browser_close to shut down the open browser instance.
The plugin bundles playwright-core as a dependency, but it requires a pre-installed Chromium-based browser on your local system. It automatically checks for common browser channels including chromium, chrome, and msedge in order, and will fall through to the next detected channel if one is missing. If no compatible browser is found, you will need to run the command npx playwright install chromium to install a compatible build, or manually set the executablePath or channel in the plugin configuration. The project is released under the open source MIT license, and runs headless by default with headed mode planned for a future update.
这是一个基于Playwright开发、专为DeepSeek Harness打造的浏览器自动化原生插件。它提供了一系列以browser_开头的工具,允许DSH代理驱动真实浏览器完成操作,包括填写表单、点击页面元素、抓取单页应用内容、截取屏幕截图供代理通过read_image读取,可直接通过DSH自带的插件命令完成安装。
它面向需要让DSH代理完成网页交互任务的开发者,典型工作流是先调用browser_open打开浏览器并导航到目标网址,之后依次调用不同工具完成点击、输入、选择、等待加载等操作,需要时可截图保存到工作区供模型读取,任务结束后调用browser_close关闭浏览器。
该插件依赖Playwright,核心已打包在包中,但需要用户系统已有Chromium内核浏览器。插件会自动检测chromium、chrome、msedge通道,若未检测到则需要用户手动执行npx playwright install chromium安装,或在配置中指定浏览器可执行路径,项目采用MIT许可证开源。
请帮我安装这个 DSH 插件。安装前先完成【兼容性检查 + 安全性检查】,检查通过再动手。
插件:dsh-browser(ben7am1n/dsh-browser)
仓库:https://github.com/ben7am1n/dsh-browser
本站详情页:https://www.yhbd.top/plugins/ben7am1n-dsh-browser/
本站登记:类型 plugin · 归类 原生 DSH 插件 · 许可证 MIT · ⭐ 7 · 最近提交 2026-08-13 · 主语言 TypeScript
按下面顺序执行,每步先把结论告诉我,再进入下一步:
【1 兼容性检查】
① 我这边:DSH 版本、Node 版本、操作系统、当前 profile(web / desktop)。
② 读它的 README、package.json、插件 manifest,列出它要求的 DSH 版本 / Node 版本 / 操作系统 / 外部依赖 / 需要另外先装的运行时。
③ 逐条比对,结论只写「满足 / 不满足 / 未知」三种;不满足的给出可行替代方案。
④ 检查是否和我已装的插件冲突:命令名重复、skill / tool 重名、端口占用、重复注册的 MCP server。
【2 安全性检查】
① 仓库可信度:和上面「本站登记」是否一致;star / fork 数、创建时间、最近提交,是否归档或长期停更。
② 安装脚本:逐行看 package.json 的 preinstall / install / postinstall,以及 install.sh、setup.ps1 之类脚本。出现 curl|bash、下载后直接执行、混淆代码、访问与插件功能无关的域名,立刻停下来告诉我,不要继续装。
③ 依赖:列出新增依赖,标出无人维护、或与知名包拼写近似的可疑包(typosquatting)。
④ 权限与副作用:它会读写哪些目录、访问哪些域名、需要哪些 DSH 权限(filesystem / network / shell / clipboard 等),以及怎么卸载和回滚。
⑤ 如果它要求 sudo / 管理员权限,或权限明显超出功能所需,先停下来问我。
【3 安装】
上面两步没有「不满足」和「高危项」时才执行;用官方推荐方式安装,不要自行提权。
【4 汇报】
用表格输出:检查项 / 结论 / 依据 / 是否需要我决策。拿不准的一律写「未知」并说明要我怎么确认——不要猜,也不要替我决定。
Send this message to DSH in your current session: it verifies compatibility and security first (answering met / not met / unknown item by item) and only installs once everything checks out — it will stop and ask you if it finds a high-risk item. The box scrolls; the copy is the full prompt. CLI install commands may not be accurate across systems, so DSH is the safer route.把上面这条消息直接发给当前会话里的 DSH:它会先核对兼容性与安全性(逐条给「满足 / 不满足 / 未知」),确认没问题再安装,有高危项会停下来问你。框内可滚动,复制到的是完整提示词;安装命令不一定准确,发给 DSH 更稳。
- Only 7 stars - very few users, little community feedback星标只有 7,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 --profile web add dsh-browser
把 ben7am1n/dsh-browser 加入你的 DSH 配置(web profile)即可启用。
READMEREADME
dsh-browser
Browser automation for DeepSeek Harness: a Playwright-powered browser_* tool family so the agent can drive real web pages — fill forms, click through flows, scrape SPA content, and take screenshots it can then view with read_image.
Install
dsh plugin --profile <name> add dsh-browser
The plugin depends on playwright-core (bundled with the package) and needs a Chromium-based browser. It tries chromium, then chrome, then msedge channels automatically; if none is found:
npx playwright install chromium
or configure launch.executablePath / launch.channel in the plugin config.
Tools
| Tool | Purpose |
|---|---|
browser_open |
Open (or reuse) the browser, optionally navigate to a URL |
browser_navigate |
Navigate the current page and wait for load |
browser_click |
Click an element by CSS selector |
browser_type |
Type text into an input (optionally press Enter) |
browser_select |
Select option(s) in a <select> |
browser_screenshot |
Capture a PNG to <workspace>/browser-screenshots/; view it with read_image |
browser_eval |
Evaluate JavaScript in the page, return JSON |
browser_get_text / browser_get_html |
Read visible text or outer HTML |
browser_wait |
Wait for slow pages / lazy content |
browser_close |
Close the browser (reopenable by any tool) |
browser_install |
Explain or verify browser availability |
The browser persists across tool calls — open once, drive many times, close when done.
Configuration
| Field | Default | Meaning |
|---|---|---|
launch.executablePath |
— | Absolute path to a browser binary (takes precedence) |
launch.channel |
auto (chromium → chrome → msedge) |
Browser channel |
launch.navigationTimeoutMs |
— | Default navigation/action timeout |
launch.viewport |
1280×800 |
Page viewport |
screenshotDir |
browser-screenshots |
Screenshot directory under the workspace |
# profile cordis.patch.yml
- id: browser
config:
launch:
channel: chrome
navigationTimeoutMs: 30000
Notes
- Runs headless by default; a headed mode may come later.
- Screenshots land in the calling agent's workspace so the model can read them with
read_image.
License
MIT
Tencent/BrowserSkill
YaoApp/yao
anbeime/skill
Jesseovo/last30days-skill-cn
bowenliang123/dsh-context
platonai/Browser4
liustack/modsearch