dsh-ego-browser
Fisfzy
DSH(DeepSeek Harness)插件:把 ego-lite 浏览器(给 AI Agent 用的 Chromium)接入 HARNESS——13 个结构化 ego_* 工具(文本语义快照、语义定位点击、表单填充、截图、CDP 控制、任务空间隔离),内置 ego 运行时,Linux + Chrome 开箱即用,无需克隆官方仓库或手动构建。
GITHUB TOPIC
11个项目包含此标签
11 个项目
GitHub Topic 精确匹配
Fisfzy
DSH(DeepSeek Harness)插件:把 ego-lite 浏览器(给 AI Agent 用的 Chromium)接入 HARNESS——13 个结构化 ego_* 工具(文本语义快照、语义定位点击、表单填充、截图、CDP 控制、任务空间隔离),内置 ego 运行时,Linux + Chrome 开箱即用,无需克隆官方仓库或手动构建。
guo6x
Give your DSH agent hands: drive a real browser (Edge/Chrome over CDP) from the chat — ref-driven clicking, per-session browsers, and a live cockpit panel. Zero runtime deps, no API key.
ben7am1n
Playwright-powered browser automation for DeepSeek Harness
maxiaovivi
Native CloakBrowser tools for DeepSeek Harness: isolated browser sessions, snapshots, interaction, screenshots, and safe Agent routing.
longmiaoo
A production-grade, DSH-native browser runtime for DeepSeek Harness with visible human-agent collaboration.
yzd6552-commits
browser-use style browser automation plugin for DeepSeek Harness: drives a dedicated Chrome instance (persistent profile) via playwright-core — fine-grained tools, autonomous tasks, scheduling, dangerous-action confirmation, captcha hand-off
ZSeven-W
DeepSeek Harness plugin that does QA: an agent explores your app like a real user and leaves evidence, then the explored path is exported as a deterministic replay scenario you run every release. Orchestrates the browser, macOS desktop, iOS and Android drivers — never a false green.
SciF-Lin
Real browser control for DeepSeek Harness: mounts Playwright MCP as native tools so the agent drives a browser you actually use.
dshworks
ego lite browser automation for dsh that remembers — recall a site's learned tools, promote a working script into a new one, hand the user the keyboard for real. 7 tools, host-only, MIT.
jiewaigongxing
DeepSeek Harness plugin: drive Jiey Browser over MCP
ZSeven-W
DeepSeek Harness plugin for browser automation, headless first. Every agent gets its own throwaway Chrome/Edge/Chromium profile, short-lived semantic references instead of CSS selectors, scoped subtree observation, deterministic high-risk refusal, and bounded evidence.