菜单
Interact with local Chrome browser session (only on explicit user approval after being asked to inspect, debug, or interact with a page open in Chrome)
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Generate autounattend.xml for Windows 11 25H2 install media. Pre-creates a local admin account, bypasses Win11 hardware checks, AutoLogons on first boot to run debloat + soft-off privacy preferences via a staged FirstLogon script, then reboots to the sign-in screen. With `-FullUnattended` also silently wipes the target disk and pins the SKU. Use when preparing Windows install USB or ISO staging media.
Generate an agent-to-agent handoff summary when context is too long and work needs to continue in a new session. Use when the user says "handoff", "交接", "context too long", "新会话继续", or when wrapping up a long session to preserve working state for the next agent.
Convert documents, URLs, and images to clean Markdown using Markitai CLI. Use when the user asks to convert files (DOCX, PDF, PPTX, XLSX, HTML, images, etc.) or URLs to Markdown, extract text from documents, fetch and convert web pages, batch-process a directory of files, or enhance Markdown output with LLM. Triggers on "convert to markdown", "extract text", "fetch this page", "markitai", or any document-to-markdown workflow.
| name | chrome-cdp |
| description | Interact with local Chrome browser session (only on explicit user approval after being asked to inspect, debug, or interact with a page open in Chrome) |
Lightweight Chrome DevTools Protocol CLI. Connects via WebSocket — no Puppeteer, works with 100+ tabs, instant connection.
chrome://inspect/#remote-debugging → toggle switchDevToolsActivePort is in a non-standard location, set CDP_PORT_FILE to its full pathscripts/cdp.mjs list # list open pages (shows target prefixes)
scripts/cdp.mjs snap <target> # accessibility tree snapshot
scripts/cdp.mjs shot <target> [file] # screenshot (prints DPR + coordinate mapping)
scripts/cdp.mjs eval <target> <expr> # evaluate JS expression
<target> is a unique targetId prefix from list (e.g. 6BE827FA). Ambiguous prefixes are rejected.
scripts/cdp.mjs html <target> [selector] # full page or element HTML
scripts/cdp.mjs nav <target> <url> # navigate and wait for load (http/https only)
scripts/cdp.mjs net <target> # resource timing entries
scripts/cdp.mjs click <target> <selector> # click by CSS selector
scripts/cdp.mjs clickxy <target> <x> <y> # click at CSS pixel coords
scripts/cdp.mjs type <target> <text> # insert text at current focus
scripts/cdp.mjs loadall <target> <selector> [ms] # click "load more" until gone
scripts/cdp.mjs evalraw <target> <method> [json] # raw CDP command passthrough
scripts/cdp.mjs open [url] # open new tab (each triggers Allow prompt)
scripts/cdp.mjs stop [target] # stop daemon(s)
shot output is native resolution (CSS px × DPR). clickxy takes CSS pixels: divide screenshot coords by DPR.snap over html for understanding page structure — much smaller output.type (not eval) for cross-origin iframes — click/clickxy to focus first, then type.eval calls when the DOM can change between them. Use stable selectors or collect all data in one eval.Based on pasky/chrome-cdp-skill by Petr Baudiš. MIT License.