[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7C0SX3THNWFAQQ1F2X5P3KS":3,"skill-same:01M7C0SX3THNWFAQQ1F2X5P3KS":201},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7C0SX3THNWFAQQ1F2X5P3KS","browsing",null,"Use when you need direct browser control - teaches Chrome DevTools Protocol for controlling existing browser sessions, multi-tab management, form automation, and content extraction via use_browser MCP tool","zak-keown\u002Fmoe","zak-keown","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F320717241?v=4","01M7BZFXME0NW6TQX0H7AZM2KS","01M7BZFXN32VR0JN5AATRD53H4",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: browsing\ndescription: Use when you need direct browser control - teaches Chrome DevTools Protocol for controlling existing browser sessions, multi-tab management, form automation, and content extraction via use_browser MCP tool\nallowed-tools: mcp__moe-glass__use_browser\n---\n\n# Browsing with Chrome Direct\n\n## Overview\n\nControl Chrome via DevTools Protocol using the `use_browser` MCP tool. Single unified interface with auto-starting Chrome.\n\n**Announce:** \"I'm using the browsing skill to control Chrome.\"\n\n## When to Use\n\n**Use this when:**\n- Controlling authenticated sessions\n- Managing multiple tabs in running browser\n- Playwright MCP unavailable or excessive\n\n**Use Playwright MCP when:**\n- Need fresh browser instances\n- Generating screenshots\u002FPDFs\n- Prefer higher-level abstractions\n\n## Auto-Capture\n\nEvery DOM action (navigate, click, type, select, eval, keyboard_press, hover, drag_drop, double_click, right_click, file_upload) automatically saves:\n- `{prefix}.png` — viewport screenshot\n- `{prefix}.md` — page content as structured markdown\n- `{prefix}.html` — full rendered DOM\n- `{prefix}-console.txt` — browser console messages\n\nFiles are saved to the session directory with sequential prefixes (001-navigate, 002-click, etc.). You must check these before using extract or screenshot actions.\n\n## The use_browser Tool\n\nSingle MCP tool with action-based interface. Chrome auto-starts on first use.\n\n**Parameters:**\n- `action` (required): Operation to perform\n- `selector` (optional): CSS or XPath selector for element operations\n- `payload` (optional): Action-specific data (string or object)\n- `timeout` (optional): Timeout in ms for await operations (default: 5000)\n\n**Active tab**: Every action operates on the current `activeTab`. Use `switch_tab` to change it.\n\n## Actions Reference\n\n### Navigation\n- **navigate**: Navigate to URL\n  - `payload`: URL string\n  - Example: `{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}`\n\n- **await_element**: Wait for element to appear\n  - `selector`: CSS selector\n  - `timeout`: Max wait time in ms\n  - Example: `{action: \"await_element\", selector: \".loaded\", timeout: 10000}`\n\n- **await_text**: Wait for text to appear\n  - `payload`: Text to wait for\n  - Example: `{action: \"await_text\", payload: \"Welcome\"}`\n\n### Interaction\n- **click**: Click element\n  - `selector`: CSS selector\n  - Example: `{action: \"click\", selector: \"button.submit\"}`\n\n- **type**: Text input\n  - `selector`: Optional — clicks to focus first\n  - `payload`: Text to type (`\\t`=Tab, `\\n`=Enter)\n  - Example: `{action: \"type\", selector: \"#email\", payload: \"user@example.com\"}`\n\n- **double_click**: Double-click element (fires dblclick event)\n  - `selector`: CSS selector\n  - Example: `{action: \"double_click\", selector: \".item\"}`\n\n- **right_click**: Right-click element (fires contextmenu event)\n  - `selector`: CSS selector\n  - Example: `{action: \"right_click\", selector: \".row\"}`\n\n- **select**: Select dropdown option\n  - `selector`: CSS selector\n  - `payload`: Option value(s)\n  - Example: `{action: \"select\", selector: \"select[name=state]\", payload: \"CA\"}`\n\n- **keyboard_press**: Press special keys (Tab, Enter, Escape, Arrow keys, F1-F12)\n  - `payload`: Key name (string) or `{\"key\": \"Tab\", \"modifiers\": {\"shift\": true, \"ctrl\": false, \"alt\": false, \"meta\": false}}`\n  - Example: `{action: \"keyboard_press\", payload: \"Tab\"}`\n  - Example with modifiers: `{action: \"keyboard_press\", payload: {\"key\": \"Tab\", \"modifiers\": {\"shift\": true}}}`\n\n### Mouse Actions (CDP-Level)\nThese use CDP Input.dispatchMouseEvent, bypassing synthetic event restrictions.\n\n- **hover**: Move mouse over element (CSS :hover, tooltips, menus)\n  - `selector`: CSS selector\n  - Example: `{action: \"hover\", selector: \".menu-trigger\"}`\n\n- **drag_drop**: Drag element to target (native drag-and-drop via CDP)\n  - `selector`: Source element\n  - `payload`: Target selector or JSON coordinates `{\"x\":N,\"y\":N}`\n  - Example: `{action: \"drag_drop\", selector: \".card\", payload: \".column-2\"}`\n\n- **mouse_move**: Move mouse to coordinates\n  - `payload`: JSON `{\"x\":N,\"y\":N}` (optional: `steps`, `fromX`, `fromY` for smooth movement)\n  - Example: `{action: \"mouse_move\", payload: \"{\\\"x\\\":100,\\\"y\\\":200}\"}`\n\n- **scroll**: Scroll via mouse wheel events\n  - `payload`: Direction (up\u002Fdown\u002Fleft\u002Fright) or JSON `{\"deltaX\":N,\"deltaY\":N}`\n  - `selector`: Optional — scroll within element\n  - Example: `{action: \"scroll\", payload: \"down\"}`\n\n### File Upload\n- **file_upload**: Set files on input[type=file] elements (can't be done via JavaScript)\n  - `selector`: File input element\n  - `payload`: File path or JSON `{\"files\":[\"\u002Fpath\u002Fa.pdf\",\"\u002Fpath\u002Fb.jpg\"]}`\n  - Example: `{action: \"file_upload\", selector: \"#upload\", payload: \"\u002Ftmp\u002Fdoc.pdf\"}`\n\n### Extraction\n- **extract**: Get page content\n  - `payload`: Format ('markdown'|'text'|'html')\n  - `selector`: Optional - limit to element\n  - Example: `{action: \"extract\", payload: \"markdown\"}`\n  - Example: `{action: \"extract\", payload: \"text\", selector: \"h1\"}`\n\n- **attr**: Get element attribute\n  - `selector`: CSS selector\n  - `payload`: Attribute name\n  - Example: `{action: \"attr\", selector: \"a.download\", payload: \"href\"}`\n\n- **eval**: Execute JavaScript\n  - `payload`: JavaScript code\n  - Example: `{action: \"eval\", payload: \"document.title\"}`\n\n### Export\n- **screenshot**: Capture screenshot of a specific element\n  - `payload`: Filename\n  - `selector`: Optional - screenshot specific element\n  - Viewport screenshots are auto-captured after every DOM action. Use this only when you need a specific element.\n  - Example: `{action: \"screenshot\", payload: \"\u002Ftmp\u002Fchart.png\", selector: \".chart\"}`\n\n### Tab Management\n- **list_tabs**: List all open tabs\n  - Example: `{action: \"list_tabs\"}`\n\n- **new_tab**: Create new tab\n  - Example: `{action: \"new_tab\"}`\n\n- **close_tab**: Close the active tab\n  - Example: `{action: \"close_tab\"}`\n\n- **switch_tab**: Switch the active tab (sticky — stays until changed)\n  - `payload`: Tab index (number), URL substring, or title substring\n  - Example: `{action: \"switch_tab\", payload: 1}` (by index)\n  - Example: `{action: \"switch_tab\", payload: \"example.com\"}` (by URL substring)\n  - Example: `{action: \"switch_tab\", payload: \"GitHub\"}` (by title substring)\n\n### Browser Mode Control\n- **show_browser**: Make browser window visible (headed mode)\n  - Example: `{action: \"show_browser\"}`\n  - ⚠️ **WARNING**: Restarts Chrome, reloads pages via GET, loses POST state\n\n- **hide_browser**: Switch to headless mode (invisible browser)\n  - Example: `{action: \"hide_browser\"}`\n  - ⚠️ **WARNING**: Restarts Chrome, reloads pages via GET, loses POST state\n\n- **browser_mode**: Check current browser mode, port, and profile\n  - Example: `{action: \"browser_mode\"}`\n  - Returns: `{\"headless\": true|false, \"mode\": \"headless\"|\"headed\", \"running\": true|false, \"port\": 9222, \"profile\": \"name\", \"profileDir\": \"\u002Fpath\"}`\n\n### Profile Management\n- **set_profile**: Change Chrome profile (must kill Chrome first)\n  - Example: `{action: \"set_profile\", \"payload\": \"browser-user\"}`\n  - ⚠️ **WARNING**: Chrome must be stopped first\n  - **Side effect**: marks the profile as explicit, opting out of auto-disambiguation (see below)\n\n- **get_profile**: Get current profile name and directory\n  - Example: `{action: \"get_profile\"}`\n  - Returns: `{\"profile\": \"name\", \"profileDir\": \"\u002Fpath\"}`\n\n**Default behavior**: Chrome starts in **headless mode** with **\"moe-glass\" profile** on a **dynamically allocated port** (range 9222-12111). Override the port with `CHROME_WS_PORT`; override the profile with `CHROME_WS_PROFILE`.\n\n**Auto-disambiguation across parallel MCPs**:\nWhen two MCP servers start on the same host with the default profile, the first claims `moe-glass` (port 9222) and later ones silently fall through to `moe-glass-2` (port 9223), `moe-glass-3`, etc. Each MCP drives its own Chrome with its own profile dir; they don't fight over `activeTab`. The bridge tracks ownership via a lock file at `~\u002F.cache\u002Fmoe\u002Fbrowser-profiles\u002F\u003Cprofile>.mcp.lock`; stale locks (dead PIDs) are reclaimed automatically.\n\nTo opt **out** of disambiguation — e.g., to intentionally share Chrome between a long-lived `chrome-ws` CLI session and your MCP — set the profile name explicitly:\n- Env var: `CHROME_WS_PROFILE=my-profile`\n- Or: `{action: \"set_profile\", payload: \"my-profile\"}` at runtime\n\nAn explicit profile name still acquires the lock, but on conflict the bridge **shares** rather than disambiguates — the second process reconnects to the first's Chrome (the original reconnect-on-restart behavior).\n\n### Chrome Lifecycle (Recovery)\n- **kill_chrome**: Kill the Chrome process this MCP is driving\n  - Example: `{action: \"kill_chrome\"}`\n  - Releases the meta.json; next page action auto-restarts Chrome\n\n- **restart_chrome**: kill_chrome + immediate spawn\n  - Example: `{action: \"restart_chrome\"}`\n\n**Auto-restart banner**: when the bridge has to spawn a fresh Chrome (because the previous one died or was killed externally — e.g., `kill -9 \u003Cpid>` from the shell), the first response after the restart prepends:\n```\n[Chrome auto-restarted; URL reset to about:blank. Re-navigate to continue.]\n```\nTreat this as a signal that your prior URL \u002F tab state is gone — re-navigate before assuming anything about the current page.\n\n### Console Logging\nCapture browser console output for the active tab. Buffer is keyed by the page session's `sessionId`, so it survives `close_tab`\u002F`new_tab` ordering quirks. Levels: `log`, `info`, `warn`, `error`.\n\n- **enable_console_logging**: Start capturing\n  - Example: `{action: \"enable_console_logging\"}`\n\n- **get_console_messages**: Read captured messages\n  - All: `{action: \"get_console_messages\"}`\n  - Since timestamp (epoch ms): `{action: \"get_console_messages\", payload: {since: 1716000000000}}`\n  - Returns: array of `{timestamp, level, text}` entries\n\n- **clear_console_messages**: Reset the buffer\n  - Example: `{action: \"clear_console_messages\"}`\n\n### Dialog Handling\nNative dialogs (JS alert\u002Fconfirm\u002Fprompt, beforeunload, HTTP basic-auth, permission prompts, device choosers) pause the page. While a dialog is open, page-targeted actions (`extract`, `click`, `eval`, etc.) return a refusal whose text contains `Page is behind a dialog` and lists the available `dialog::*` selectors.\n\nWhen a dialog fires **during** a `navigate` (typical for HTTP basic-auth), `navigate` itself throws with the dialog grammar in the message — you don't have to issue a separate page-targeted call to discover the dialog.\n\nHandle dialogs by clicking\u002Ftyping a `dialog::*` selector:\n- `{action: \"click\", selector: \"dialog::accept\"}` — accept JS alert\u002Fconfirm\u002Fprompt, beforeunload, permission grant\n- `{action: \"click\", selector: \"dialog::dismiss\"}` — dismiss \u002F cancel \u002F deny\n- `{action: \"type\", selector: \"dialog::prompt\", payload: \"text\"}` then accept — respond to JS prompt\n- `{action: \"type\", selector: \"dialog::username\", payload: \"alice\"}` + `{action: \"type\", selector: \"dialog::password\", payload: \"secret\"}` + `{action: \"click\", selector: \"dialog::accept\"}` — HTTP basic-auth\n- `{action: \"click\", selector: \"dialog::device[id=\\\"\u003CdeviceId>\\\"]\"}` — pick a WebUSB\u002FBluetooth\u002FSerial\u002FHID device\n\n**Critical caveats when toggling modes**:\n1. **Chrome must restart** - Cannot switch headless\u002Fheaded mode on running Chrome\n2. **Pages reload via GET** - All open tabs are reopened with GET requests\n3. **POST state is lost** - Form submissions, POST results, and POST-based navigation will be lost\n4. **Session state is lost** - Any client-side state (JavaScript variables, etc.) is cleared\n5. **Cookies\u002Fauth may persist** - Uses same user data directory, so logged-in sessions may survive\n\n**When to use headed mode**:\n- Debugging visual rendering issues\n- Demonstrating browser behavior to user\n- Testing features that only work with visible browser\n- Debugging issues that don't reproduce in headless mode\n\n**When to stay in headless mode** (default):\n- All other cases - faster, cleaner, less intrusive\n- Screenshots work perfectly in headless mode\n- Most automation works identically in both modes\n\n**Profile management**:\nProfiles store persistent browser data (cookies, localStorage, extensions, auth sessions).\n\n**Profile locations**:\n- macOS: `~\u002FLibrary\u002FCaches\u002Fmoe\u002Fbrowser-profiles\u002F{name}\u002F`\n- Linux: `~\u002F.cache\u002Fmoe\u002Fbrowser-profiles\u002F{name}\u002F`\n- Windows: `%LOCALAPPDATA%\u002Fmoe\u002Fbrowser-profiles\u002F{name}\u002F`\n\n**When to use separate profiles**:\n- **Default profile (\"moe-glass\")**: General automation, shared sessions\n- **Agent-specific profiles**: Isolate different agents' browser state\n  - Example: browser-user agent uses \"browser-user\" profile\n- **Task-specific profiles**: Testing with different user contexts\n  - Example: \"test-logged-in\" vs \"test-logged-out\"\n\n**Profile data persists across**:\n- Chrome restarts\n- Mode toggles (headless ↔ headed)\n- System reboots (data is in cache directory)\n\n**To use a different profile**:\n1. Kill Chrome if running: `await chromeLib.killChrome()`\n2. Set profile: `{action: \"set_profile\", \"payload\": \"my-profile\"}`\n3. Start Chrome: Next navigate\u002Faction will use new profile\n\n## Quick Start Pattern\n\n```\nNavigate and extract:\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"await_element\", selector: \"h1\"}\n{action: \"extract\", payload: \"text\", selector: \"h1\"}\n```\n\n## Common Patterns\n\n### Fill and Submit Form\n```\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\u002Flogin\"}\n{action: \"await_element\", selector: \"input[name=email]\"}\n{action: \"type\", selector: \"input[name=email]\", payload: \"user@example.com\"}\n{action: \"type\", selector: \"input[name=password]\", payload: \"pass123\"}\n{action: \"keyboard_press\", payload: \"Enter\"}\n{action: \"await_text\", payload: \"Welcome\"}\n```\n\nUses `keyboard_press` to submit the form.\n\n### Multi-Tab Workflow\n```\n{action: \"list_tabs\"}\n{action: \"switch_tab\", payload: 2}\n{action: \"click\", selector: \"a.email\"}\n{action: \"await_element\", selector: \".content\"}\n{action: \"extract\", payload: \"text\", selector: \".amount\"}\n```\n\n### Dynamic Content\n```\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"type\", selector: \"input[name=q]\", payload: \"query\"}\n{action: \"click\", selector: \"button.search\"}\n{action: \"await_element\", selector: \".results\"}\n{action: \"extract\", payload: \"text\", selector: \".result-title\"}\n```\n\n### Get Link Attribute\n```\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"await_element\", selector: \"a.download\"}\n{action: \"attr\", selector: \"a.download\", payload: \"href\"}\n```\n\n### Execute JavaScript\n```\n{action: \"eval\", payload: \"document.querySelectorAll('a').length\"}\n{action: \"eval\", payload: \"Array.from(document.querySelectorAll('a')).map(a => a.href)\"}\n```\n\n### Resize Viewport (Responsive Testing)\nUse `eval` to resize the browser window for testing responsive layouts:\n```\n{action: \"eval\", payload: \"window.resizeTo(375, 812); 'Resized to mobile'\"}\n{action: \"eval\", payload: \"window.resizeTo(768, 1024); 'Resized to tablet'\"}\n{action: \"eval\", payload: \"window.resizeTo(1920, 1080); 'Resized to desktop'\"}\n```\n\n**Note**: This resizes the window, not device emulation. It won't change:\n- Device pixel ratio (retina displays)\n- Touch events\n- User-Agent string\n\nFor most responsive testing, window resize is sufficient.\n\n### Clear Cookies\nUse `eval` to clear cookies accessible to JavaScript:\n```\n{action: \"eval\", payload: \"document.cookie.split(';').forEach(c => { document.cookie = c.trim().split('=')[0] + '=;expires=Thu, 01 Jan 1970 00:00:00 GMT;path=\u002F'; }); 'Cookies cleared'\"}\n```\n\n**Note**: This clears cookies accessible to JavaScript. It won't clear:\n- httpOnly cookies (server-side only)\n- Cookies from other domains\n\nFor most logout\u002Freset scenarios, this is sufficient.\n\n### Scroll Page\n```\n{action: \"scroll\", payload: \"down\"}\n{action: \"scroll\", payload: \"up\"}\n{action: \"scroll\", selector: \".container\", payload: \"{\\\"deltaX\\\":0,\\\"deltaY\\\":500}\"}\n```\n\nUses real mouse wheel events (vs `eval` + `scrollTo` which bot detectors flag).\n\n## Tips\n\n**Always wait before interaction:**\nDon't click or fill immediately after navigate - pages need time to load.\n\n```\n\u002F\u002F BAD - might fail if page slow\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"click\", selector: \"button\"}  \u002F\u002F May fail!\n\n\u002F\u002F GOOD - wait first\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"await_element\", selector: \"button\"}\n{action: \"click\", selector: \"button\"}\n```\n\n**Use specific selectors:**\nAvoid generic selectors that match multiple elements.\n\n```\n\u002F\u002F BAD - matches first button\n{action: \"click\", selector: \"button\"}\n\n\u002F\u002F GOOD - specific\n{action: \"click\", selector: \"button[type=submit]\"}\n{action: \"click\", selector: \"#login-button\"}\n```\n\n**Submit forms:**\nUse `keyboard_press` with Enter after `type`, or append `\\n` to the payload.\n\n```\n{action: \"type\", selector: \"#search\", payload: \"query\"}\n{action: \"keyboard_press\", payload: \"Enter\"}\n```\n\n**Check content first:**\nExtract page content to verify selectors before building workflow.\n\n```\n{action: \"extract\", payload: \"html\"}\n```\n\n## Troubleshooting\n\n**Element not found:**\n- Use `await_element` before interaction\n- Verify selector with `extract` action using 'html' format\n\n**Timeout errors:**\n- Increase timeout: `{timeout: 30000}` for slow pages\n- Wait for specific element instead of text\n\n**Wrong tab active:**\n- Use `list_tabs` to see all open tabs\n- Use `switch_tab` with a URL or title substring to reliably switch tabs\n- Tab indices shift when tabs close — prefer URL\u002Ftitle-based switching\n\n**eval returns `[object Object]`:**\n- Use `JSON.stringify()` for complex objects: `{action: \"eval\", payload: \"JSON.stringify({name: 'test'})\"}`\n- For async functions: `{action: \"eval\", payload: \"JSON.stringify(await yourAsyncFunction())\"}`\n\n## Test Automation (Advanced)\n\n\u003Cdetails>\n\u003Csummary>Click to expand test automation guidance\u003C\u002Fsummary>\n\nWhen building test automation, you have two approaches:\n\n### Approach 1: use_browser MCP (Simple Tests)\nBest for: Single-step tests, direct Claude control during conversation\n\n```json\n{\"action\": \"navigate\", \"payload\": \"https:\u002F\u002Fapp.com\"}\n{\"action\": \"click\", \"selector\": \"#test-button\"}\n{\"action\": \"eval\", \"payload\": \"JSON.stringify({passed: document.querySelector('.success') !== null})\"}\n```\n\n### Approach 2: chrome-ws CLI (Complex Tests)\nBest for: Multi-step test suites, standalone automation scripts\n\n**Key insight**: `chrome-ws` is the reference implementation showing proper Chrome DevTools Protocol usage. When `use_browser` doesn't work as expected, examine how `chrome-ws` handles the same operation.\n\n```bash\n# Example: Automated form testing\nnode \"$SKILL\u002Fchrome-ws.mjs\" navigate 0 \"https:\u002F\u002Fapp.com\u002Fform\"\nnode \"$SKILL\u002Fchrome-ws.mjs\" fill 0 \"#email\" \"test@example.com\"\nnode \"$SKILL\u002Fchrome-ws.mjs\" click 0 \"button[type=submit]\"\nnode \"$SKILL\u002Fchrome-ws.mjs\" wait-text 0 \"Success\"\n```\n\n### When use_browser Fails\n1. **Check chrome-ws source code** - It shows the correct CDP pattern\n2. **Use chrome-ws to verify** - Test the same operation via CLI\n3. **Adapt the pattern** - Apply the working CDP approach to use_browser\n\n### Common Test Automation Patterns\n- **Form validation**: Fill forms, check error states\n- **UI state testing**: Click elements, verify DOM changes\n- **Performance testing**: Measure load times, capture metrics\n- **Screenshot comparison**: Capture before\u002Fafter states\n\n\u003C\u002Fdetails>\n\n## Advanced Usage\n\nFor command-line usage outside Claude Code, see [COMMANDLINE-USAGE.md](COMMANDLINE-USAGE.md).\n\nFor detailed examples, see [EXAMPLES.md](EXAMPLES.md).\n\n## Protocol Reference\n\nFull CDP documentation: https:\u002F\u002Fchromedevtools.github.io\u002Fdevtools-protocol\u002F\n","# Browsing with Chrome Direct\n\n## Overview\n\nControl Chrome via DevTools Protocol using the `use_browser` MCP tool. Single unified interface with auto-starting Chrome.\n\n**Announce:** \"I'm using the browsing skill to control Chrome.\"\n\n## When to Use\n\n**Use this when:**\n- Controlling authenticated sessions\n- Managing multiple tabs in running browser\n- Playwright MCP unavailable or excessive\n\n**Use Playwright MCP when:**\n- Need fresh browser instances\n- Generating screenshots\u002FPDFs\n- Prefer higher-level abstractions\n\n## Auto-Capture\n\nEvery DOM action (navigate, click, type, select, eval, keyboard_press, hover, drag_drop, double_click, right_click, file_upload) automatically saves:\n- `{prefix}.png` — viewport screenshot\n- `{prefix}.md` — page content as structured markdown\n- `{prefix}.html` — full rendered DOM\n- `{prefix}-console.txt` — browser console messages\n\nFiles are saved to the session directory with sequential prefixes (001-navigate, 002-click, etc.). You must check these before using extract or screenshot actions.\n\n## The use_browser Tool\n\nSingle MCP tool with action-based interface. Chrome auto-starts on first use.\n\n**Parameters:**\n- `action` (required): Operation to perform\n- `selector` (optional): CSS or XPath selector for element operations\n- `payload` (optional): Action-specific data (string or object)\n- `timeout` (optional): Timeout in ms for await operations (default: 5000)\n\n**Active tab**: Every action operates on the current `activeTab`. Use `switch_tab` to change it.\n\n## Actions Reference\n\n### Navigation\n- **navigate**: Navigate to URL\n  - `payload`: URL string\n  - Example: `{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}`\n\n- **await_element**: Wait for element to appear\n  - `selector`: CSS selector\n  - `timeout`: Max wait time in ms\n  - Example: `{action: \"await_element\", selector: \".loaded\", timeout: 10000}`\n\n- **await_text**: Wait for text to appear\n  - `payload`: Text to wait for\n  - Example: `{action: \"await_text\", payload: \"Welcome\"}`\n\n### Interaction\n- **click**: Click element\n  - `selector`: CSS selector\n  - Example: `{action: \"click\", selector: \"button.submit\"}`\n\n- **type**: Text input\n  - `selector`: Optional — clicks to focus first\n  - `payload`: Text to type (`\\t`=Tab, `\\n`=Enter)\n  - Example: `{action: \"type\", selector: \"#email\", payload: \"user@example.com\"}`\n\n- **double_click**: Double-click element (fires dblclick event)\n  - `selector`: CSS selector\n  - Example: `{action: \"double_click\", selector: \".item\"}`\n\n- **right_click**: Right-click element (fires contextmenu event)\n  - `selector`: CSS selector\n  - Example: `{action: \"right_click\", selector: \".row\"}`\n\n- **select**: Select dropdown option\n  - `selector`: CSS selector\n  - `payload`: Option value(s)\n  - Example: `{action: \"select\", selector: \"select[name=state]\", payload: \"CA\"}`\n\n- **keyboard_press**: Press special keys (Tab, Enter, Escape, Arrow keys, F1-F12)\n  - `payload`: Key name (string) or `{\"key\": \"Tab\", \"modifiers\": {\"shift\": true, \"ctrl\": false, \"alt\": false, \"meta\": false}}`\n  - Example: `{action: \"keyboard_press\", payload: \"Tab\"}`\n  - Example with modifiers: `{action: \"keyboard_press\", payload: {\"key\": \"Tab\", \"modifiers\": {\"shift\": true}}}`\n\n### Mouse Actions (CDP-Level)\nThese use CDP Input.dispatchMouseEvent, bypassing synthetic event restrictions.\n\n- **hover**: Move mouse over element (CSS :hover, tooltips, menus)\n  - `selector`: CSS selector\n  - Example: `{action: \"hover\", selector: \".menu-trigger\"}`\n\n- **drag_drop**: Drag element to target (native drag-and-drop via CDP)\n  - `selector`: Source element\n  - `payload`: Target selector or JSON coordinates `{\"x\":N,\"y\":N}`\n  - Example: `{action: \"drag_drop\", selector: \".card\", payload: \".column-2\"}`\n\n- **mouse_move**: Move mouse to coordinates\n  - `payload`: JSON `{\"x\":N,\"y\":N}` (optional: `steps`, `fromX`, `fromY` for smooth movement)\n  - Example: `{action: \"mouse_move\", payload: \"{\\\"x\\\":100,\\\"y\\\":200}\"}`\n\n- **scroll**: Scroll via mouse wheel events\n  - `payload`: Direction (up\u002Fdown\u002Fleft\u002Fright) or JSON `{\"deltaX\":N,\"deltaY\":N}`\n  - `selector`: Optional — scroll within element\n  - Example: `{action: \"scroll\", payload: \"down\"}`\n\n### File Upload\n- **file_upload**: Set files on input[type=file] elements (can't be done via JavaScript)\n  - `selector`: File input element\n  - `payload`: File path or JSON `{\"files\":[\"\u002Fpath\u002Fa.pdf\",\"\u002Fpath\u002Fb.jpg\"]}`\n  - Example: `{action: \"file_upload\", selector: \"#upload\", payload: \"\u002Ftmp\u002Fdoc.pdf\"}`\n\n### Extraction\n- **extract**: Get page content\n  - `payload`: Format ('markdown'|'text'|'html')\n  - `selector`: Optional - limit to element\n  - Example: `{action: \"extract\", payload: \"markdown\"}`\n  - Example: `{action: \"extract\", payload: \"text\", selector: \"h1\"}`\n\n- **attr**: Get element attribute\n  - `selector`: CSS selector\n  - `payload`: Attribute name\n  - Example: `{action: \"attr\", selector: \"a.download\", payload: \"href\"}`\n\n- **eval**: Execute JavaScript\n  - `payload`: JavaScript code\n  - Example: `{action: \"eval\", payload: \"document.title\"}`\n\n### Export\n- **screenshot**: Capture screenshot of a specific element\n  - `payload`: Filename\n  - `selector`: Optional - screenshot specific element\n  - Viewport screenshots are auto-captured after every DOM action. Use this only when you need a specific element.\n  - Example: `{action: \"screenshot\", payload: \"\u002Ftmp\u002Fchart.png\", selector: \".chart\"}`\n\n### Tab Management\n- **list_tabs**: List all open tabs\n  - Example: `{action: \"list_tabs\"}`\n\n- **new_tab**: Create new tab\n  - Example: `{action: \"new_tab\"}`\n\n- **close_tab**: Close the active tab\n  - Example: `{action: \"close_tab\"}`\n\n- **switch_tab**: Switch the active tab (sticky — stays until changed)\n  - `payload`: Tab index (number), URL substring, or title substring\n  - Example: `{action: \"switch_tab\", payload: 1}` (by index)\n  - Example: `{action: \"switch_tab\", payload: \"example.com\"}` (by URL substring)\n  - Example: `{action: \"switch_tab\", payload: \"GitHub\"}` (by title substring)\n\n### Browser Mode Control\n- **show_browser**: Make browser window visible (headed mode)\n  - Example: `{action: \"show_browser\"}`\n  - ⚠️ **WARNING**: Restarts Chrome, reloads pages via GET, loses POST state\n\n- **hide_browser**: Switch to headless mode (invisible browser)\n  - Example: `{action: \"hide_browser\"}`\n  - ⚠️ **WARNING**: Restarts Chrome, reloads pages via GET, loses POST state\n\n- **browser_mode**: Check current browser mode, port, and profile\n  - Example: `{action: \"browser_mode\"}`\n  - Returns: `{\"headless\": true|false, \"mode\": \"headless\"|\"headed\", \"running\": true|false, \"port\": 9222, \"profile\": \"name\", \"profileDir\": \"\u002Fpath\"}`\n\n### Profile Management\n- **set_profile**: Change Chrome profile (must kill Chrome first)\n  - Example: `{action: \"set_profile\", \"payload\": \"browser-user\"}`\n  - ⚠️ **WARNING**: Chrome must be stopped first\n  - **Side effect**: marks the profile as explicit, opting out of auto-disambiguation (see below)\n\n- **get_profile**: Get current profile name and directory\n  - Example: `{action: \"get_profile\"}`\n  - Returns: `{\"profile\": \"name\", \"profileDir\": \"\u002Fpath\"}`\n\n**Default behavior**: Chrome starts in **headless mode** with **\"moe-glass\" profile** on a **dynamically allocated port** (range 9222-12111). Override the port with `CHROME_WS_PORT`; override the profile with `CHROME_WS_PROFILE`.\n\n**Auto-disambiguation across parallel MCPs**:\nWhen two MCP servers start on the same host with the default profile, the first claims `moe-glass` (port 9222) and later ones silently fall through to `moe-glass-2` (port 9223), `moe-glass-3`, etc. Each MCP drives its own Chrome with its own profile dir; they don't fight over `activeTab`. The bridge tracks ownership via a lock file at `~\u002F.cache\u002Fmoe\u002Fbrowser-profiles\u002F\u003Cprofile>.mcp.lock`; stale locks (dead PIDs) are reclaimed automatically.\n\nTo opt **out** of disambiguation — e.g., to intentionally share Chrome between a long-lived `chrome-ws` CLI session and your MCP — set the profile name explicitly:\n- Env var: `CHROME_WS_PROFILE=my-profile`\n- Or: `{action: \"set_profile\", payload: \"my-profile\"}` at runtime\n\nAn explicit profile name still acquires the lock, but on conflict the bridge **shares** rather than disambiguates — the second process reconnects to the first's Chrome (the original reconnect-on-restart behavior).\n\n### Chrome Lifecycle (Recovery)\n- **kill_chrome**: Kill the Chrome process this MCP is driving\n  - Example: `{action: \"kill_chrome\"}`\n  - Releases the meta.json; next page action auto-restarts Chrome\n\n- **restart_chrome**: kill_chrome + immediate spawn\n  - Example: `{action: \"restart_chrome\"}`\n\n**Auto-restart banner**: when the bridge has to spawn a fresh Chrome (because the previous one died or was killed externally — e.g., `kill -9 \u003Cpid>` from the shell), the first response after the restart prepends:\n```\n[Chrome auto-restarted; URL reset to about:blank. Re-navigate to continue.]\n```\nTreat this as a signal that your prior URL \u002F tab state is gone — re-navigate before assuming anything about the current page.\n\n### Console Logging\nCapture browser console output for the active tab. Buffer is keyed by the page session's `sessionId`, so it survives `close_tab`\u002F`new_tab` ordering quirks. Levels: `log`, `info`, `warn`, `error`.\n\n- **enable_console_logging**: Start capturing\n  - Example: `{action: \"enable_console_logging\"}`\n\n- **get_console_messages**: Read captured messages\n  - All: `{action: \"get_console_messages\"}`\n  - Since timestamp (epoch ms): `{action: \"get_console_messages\", payload: {since: 1716000000000}}`\n  - Returns: array of `{timestamp, level, text}` entries\n\n- **clear_console_messages**: Reset the buffer\n  - Example: `{action: \"clear_console_messages\"}`\n\n### Dialog Handling\nNative dialogs (JS alert\u002Fconfirm\u002Fprompt, beforeunload, HTTP basic-auth, permission prompts, device choosers) pause the page. While a dialog is open, page-targeted actions (`extract`, `click`, `eval`, etc.) return a refusal whose text contains `Page is behind a dialog` and lists the available `dialog::*` selectors.\n\nWhen a dialog fires **during** a `navigate` (typical for HTTP basic-auth), `navigate` itself throws with the dialog grammar in the message — you don't have to issue a separate page-targeted call to discover the dialog.\n\nHandle dialogs by clicking\u002Ftyping a `dialog::*` selector:\n- `{action: \"click\", selector: \"dialog::accept\"}` — accept JS alert\u002Fconfirm\u002Fprompt, beforeunload, permission grant\n- `{action: \"click\", selector: \"dialog::dismiss\"}` — dismiss \u002F cancel \u002F deny\n- `{action: \"type\", selector: \"dialog::prompt\", payload: \"text\"}` then accept — respond to JS prompt\n- `{action: \"type\", selector: \"dialog::username\", payload: \"alice\"}` + `{action: \"type\", selector: \"dialog::password\", payload: \"secret\"}` + `{action: \"click\", selector: \"dialog::accept\"}` — HTTP basic-auth\n- `{action: \"click\", selector: \"dialog::device[id=\\\"\u003CdeviceId>\\\"]\"}` — pick a WebUSB\u002FBluetooth\u002FSerial\u002FHID device\n\n**Critical caveats when toggling modes**:\n1. **Chrome must restart** - Cannot switch headless\u002Fheaded mode on running Chrome\n2. **Pages reload via GET** - All open tabs are reopened with GET requests\n3. **POST state is lost** - Form submissions, POST results, and POST-based navigation will be lost\n4. **Session state is lost** - Any client-side state (JavaScript variables, etc.) is cleared\n5. **Cookies\u002Fauth may persist** - Uses same user data directory, so logged-in sessions may survive\n\n**When to use headed mode**:\n- Debugging visual rendering issues\n- Demonstrating browser behavior to user\n- Testing features that only work with visible browser\n- Debugging issues that don't reproduce in headless mode\n\n**When to stay in headless mode** (default):\n- All other cases - faster, cleaner, less intrusive\n- Screenshots work perfectly in headless mode\n- Most automation works identically in both modes\n\n**Profile management**:\nProfiles store persistent browser data (cookies, localStorage, extensions, auth sessions).\n\n**Profile locations**:\n- macOS: `~\u002FLibrary\u002FCaches\u002Fmoe\u002Fbrowser-profiles\u002F{name}\u002F`\n- Linux: `~\u002F.cache\u002Fmoe\u002Fbrowser-profiles\u002F{name}\u002F`\n- Windows: `%LOCALAPPDATA%\u002Fmoe\u002Fbrowser-profiles\u002F{name}\u002F`\n\n**When to use separate profiles**:\n- **Default profile (\"moe-glass\")**: General automation, shared sessions\n- **Agent-specific profiles**: Isolate different agents' browser state\n  - Example: browser-user agent uses \"browser-user\" profile\n- **Task-specific profiles**: Testing with different user contexts\n  - Example: \"test-logged-in\" vs \"test-logged-out\"\n\n**Profile data persists across**:\n- Chrome restarts\n- Mode toggles (headless ↔ headed)\n- System reboots (data is in cache directory)\n\n**To use a different profile**:\n1. Kill Chrome if running: `await chromeLib.killChrome()`\n2. Set profile: `{action: \"set_profile\", \"payload\": \"my-profile\"}`\n3. Start Chrome: Next navigate\u002Faction will use new profile\n\n## Quick Start Pattern\n\n```\nNavigate and extract:\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"await_element\", selector: \"h1\"}\n{action: \"extract\", payload: \"text\", selector: \"h1\"}\n```\n\n## Common Patterns\n\n### Fill and Submit Form\n```\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\u002Flogin\"}\n{action: \"await_element\", selector: \"input[name=email]\"}\n{action: \"type\", selector: \"input[name=email]\", payload: \"user@example.com\"}\n{action: \"type\", selector: \"input[name=password]\", payload: \"pass123\"}\n{action: \"keyboard_press\", payload: \"Enter\"}\n{action: \"await_text\", payload: \"Welcome\"}\n```\n\nUses `keyboard_press` to submit the form.\n\n### Multi-Tab Workflow\n```\n{action: \"list_tabs\"}\n{action: \"switch_tab\", payload: 2}\n{action: \"click\", selector: \"a.email\"}\n{action: \"await_element\", selector: \".content\"}\n{action: \"extract\", payload: \"text\", selector: \".amount\"}\n```\n\n### Dynamic Content\n```\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"type\", selector: \"input[name=q]\", payload: \"query\"}\n{action: \"click\", selector: \"button.search\"}\n{action: \"await_element\", selector: \".results\"}\n{action: \"extract\", payload: \"text\", selector: \".result-title\"}\n```\n\n### Get Link Attribute\n```\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"await_element\", selector: \"a.download\"}\n{action: \"attr\", selector: \"a.download\", payload: \"href\"}\n```\n\n### Execute JavaScript\n```\n{action: \"eval\", payload: \"document.querySelectorAll('a').length\"}\n{action: \"eval\", payload: \"Array.from(document.querySelectorAll('a')).map(a => a.href)\"}\n```\n\n### Resize Viewport (Responsive Testing)\nUse `eval` to resize the browser window for testing responsive layouts:\n```\n{action: \"eval\", payload: \"window.resizeTo(375, 812); 'Resized to mobile'\"}\n{action: \"eval\", payload: \"window.resizeTo(768, 1024); 'Resized to tablet'\"}\n{action: \"eval\", payload: \"window.resizeTo(1920, 1080); 'Resized to desktop'\"}\n```\n\n**Note**: This resizes the window, not device emulation. It won't change:\n- Device pixel ratio (retina displays)\n- Touch events\n- User-Agent string\n\nFor most responsive testing, window resize is sufficient.\n\n### Clear Cookies\nUse `eval` to clear cookies accessible to JavaScript:\n```\n{action: \"eval\", payload: \"document.cookie.split(';').forEach(c => { document.cookie = c.trim().split('=')[0] + '=;expires=Thu, 01 Jan 1970 00:00:00 GMT;path=\u002F'; }); 'Cookies cleared'\"}\n```\n\n**Note**: This clears cookies accessible to JavaScript. It won't clear:\n- httpOnly cookies (server-side only)\n- Cookies from other domains\n\nFor most logout\u002Freset scenarios, this is sufficient.\n\n### Scroll Page\n```\n{action: \"scroll\", payload: \"down\"}\n{action: \"scroll\", payload: \"up\"}\n{action: \"scroll\", selector: \".container\", payload: \"{\\\"deltaX\\\":0,\\\"deltaY\\\":500}\"}\n```\n\nUses real mouse wheel events (vs `eval` + `scrollTo` which bot detectors flag).\n\n## Tips\n\n**Always wait before interaction:**\nDon't click or fill immediately after navigate - pages need time to load.\n\n```\n\u002F\u002F BAD - might fail if page slow\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"click\", selector: \"button\"}  \u002F\u002F May fail!\n\n\u002F\u002F GOOD - wait first\n{action: \"navigate\", payload: \"https:\u002F\u002Fexample.com\"}\n{action: \"await_element\", selector: \"button\"}\n{action: \"click\", selector: \"button\"}\n```\n\n**Use specific selectors:**\nAvoid generic selectors that match multiple elements.\n\n```\n\u002F\u002F BAD - matches first button\n{action: \"click\", selector: \"button\"}\n\n\u002F\u002F GOOD - specific\n{action: \"click\", selector: \"button[type=submit]\"}\n{action: \"click\", selector: \"#login-button\"}\n```\n\n**Submit forms:**\nUse `keyboard_press` with Enter after `type`, or append `\\n` to the payload.\n\n```\n{action: \"type\", selector: \"#search\", payload: \"query\"}\n{action: \"keyboard_press\", payload: \"Enter\"}\n```\n\n**Check content first:**\nExtract page content to verify selectors before building workflow.\n\n```\n{action: \"extract\", payload: \"html\"}\n```\n\n## Troubleshooting\n\n**Element not found:**\n- Use `await_element` before interaction\n- Verify selector with `extract` action using 'html' format\n\n**Timeout errors:**\n- Increase timeout: `{timeout: 30000}` for slow pages\n- Wait for specific element instead of text\n\n**Wrong tab active:**\n- Use `list_tabs` to see all open tabs\n- Use `switch_tab` with a URL or title substring to reliably switch tabs\n- Tab indices shift when tabs close — prefer URL\u002Ftitle-based switching\n\n**eval returns `[object Object]`:**\n- Use `JSON.stringify()` for complex objects: `{action: \"eval\", payload: \"JSON.stringify({name: 'test'})\"}`\n- For async functions: `{action: \"eval\", payload: \"JSON.stringify(await yourAsyncFunction())\"}`\n\n## Test Automation (Advanced)\n\n\u003Cdetails>\n\u003Csummary>Click to expand test automation guidance\u003C\u002Fsummary>\n\nWhen building test automation, you have two approaches:\n\n### Approach 1: use_browser MCP (Simple Tests)\nBest for: Single-step tests, direct Claude control during conversation\n\n```json\n{\"action\": \"navigate\", \"payload\": \"https:\u002F\u002Fapp.com\"}\n{\"action\": \"click\", \"selector\": \"#test-button\"}\n{\"action\": \"eval\", \"payload\": \"JSON.stringify({passed: document.querySelector('.success') !== null})\"}\n```\n\n### Approach 2: chrome-ws CLI (Complex Tests)\nBest for: Multi-step test suites, standalone automation scripts\n\n**Key insight**: `chrome-ws` is the reference implementation showing proper Chrome DevTools Protocol usage. When `use_browser` doesn't work as expected, examine how `chrome-ws` handles the same operation.\n\n```bash\n# Example: Automated form testing\nnode \"$SKILL\u002Fchrome-ws.mjs\" navigate 0 \"https:\u002F\u002Fapp.com\u002Fform\"\nnode \"$SKILL\u002Fchrome-ws.mjs\" fill 0 \"#email\" \"test@example.com\"\nnode \"$SKILL\u002Fchrome-ws.mjs\" click 0 \"button[type=submit]\"\nnode \"$SKILL\u002Fchrome-ws.mjs\" wait-text 0 \"Success\"\n```\n\n### When use_browser Fails\n1. **Check chrome-ws source code** - It shows the correct CDP pattern\n2. **Use chrome-ws to verify** - Test the same operation via CLI\n3. **Adapt the pattern** - Apply the working CDP approach to use_browser\n\n### Common Test Automation Patterns\n- **Form validation**: Fill forms, check error states\n- **UI state testing**: Click elements, verify DOM changes\n- **Performance testing**: Measure load times, capture metrics\n- **Screenshot comparison**: Capture before\u002Fafter states\n\n\u003C\u002Fdetails>\n\n## Advanced Usage\n\nFor command-line usage outside Claude Code, see [COMMANDLINE-USAGE.md](COMMANDLINE-USAGE.md).\n\nFor detailed examples, see [EXAMPLES.md](EXAMPLES.md).\n\n## Protocol Reference\n\nFull CDP documentation: https:\u002F\u002Fchromedevtools.github.io\u002Fdevtools-protocol\u002F\n",{"name":5,"description":7,"allowed-tools":23},"mcp__moe-glass__use_browser","main","728894f3f85fbd4e788f015ddfd72bb6155d5dd5","plugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing",1,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe","https:\u002F\u002Fgithub.com\u002Fzak-keown","2026-09-05 06:29:06",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add zak-keown\u002Fmoe --skill browsing",{"tool":37,"command":38},"bunx","bunx skills add zak-keown\u002Fmoe --skill browsing",{"tool":40,"command":41},"pnpm","pnpm dlx skills add zak-keown\u002Fmoe --skill browsing","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002FSKILL.md",true,[45,51,55,59,65,69,73,77,81,85,89,93,97,101,105,109,113,117,121,125,129,133,137,141,145,149,153,157,161,165,169,173,177,181,185,189,193,197],{"path":46,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":49,"sourceHtmlUrl":50},"COMMANDLINE-USAGE.md","REFERENCE","参考文档",14878,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002FCOMMANDLINE-USAGE.md",{"path":52,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":53,"sourceHtmlUrl":54},"EXAMPLES.md",20505,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002FEXAMPLES.md",{"path":56,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":57,"sourceHtmlUrl":58},"README.md",1281,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002FREADME.md",{"path":60,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":63,"sourceHtmlUrl":64},"scripts\u002Fchrome-ws-lib.mjs","SCRIPT","脚本",9222,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Fchrome-ws-lib.mjs",{"path":66,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":67,"sourceHtmlUrl":68},"scripts\u002Fchrome-ws.mjs",30997,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Fchrome-ws.mjs",{"path":70,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":71,"sourceHtmlUrl":72},"scripts\u002Fhost-override.mjs",1888,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Fhost-override.mjs",{"path":74,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":75,"sourceHtmlUrl":76},"scripts\u002Flib\u002Fbrowser-bridge.mjs",4299,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fbrowser-bridge.mjs",{"path":78,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":79,"sourceHtmlUrl":80},"scripts\u002Flib\u002Fbrowser-session.mjs",3716,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fbrowser-session.mjs",{"path":82,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":83,"sourceHtmlUrl":84},"scripts\u002Flib\u002Fcapture.mjs",14347,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fcapture.mjs",{"path":86,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":87,"sourceHtmlUrl":88},"scripts\u002Flib\u002Fcdp-router.mjs",1692,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fcdp-router.mjs",{"path":90,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":91,"sourceHtmlUrl":92},"scripts\u002Flib\u002Fcdp-utils.mjs",313,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fcdp-utils.mjs",{"path":94,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":95,"sourceHtmlUrl":96},"scripts\u002Flib\u002Fchrome-launcher-helpers.mjs",8661,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fchrome-launcher-helpers.mjs",{"path":98,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":99,"sourceHtmlUrl":100},"scripts\u002Flib\u002Fchrome-process.mjs",11655,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fchrome-process.mjs",{"path":102,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":103,"sourceHtmlUrl":104},"scripts\u002Flib\u002Fconsole-logging.mjs",1927,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fconsole-logging.mjs",{"path":106,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":107,"sourceHtmlUrl":108},"scripts\u002Flib\u002Fcookies.mjs",263,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fcookies.mjs",{"path":110,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":111,"sourceHtmlUrl":112},"scripts\u002Flib\u002Fdialogs-render.mjs",5321,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fdialogs-render.mjs",{"path":114,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":115,"sourceHtmlUrl":116},"scripts\u002Flib\u002Fdialogs-router.mjs",4574,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fdialogs-router.mjs",{"path":118,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":119,"sourceHtmlUrl":120},"scripts\u002Flib\u002Fdialogs.mjs",8164,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fdialogs.mjs",{"path":122,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":123,"sourceHtmlUrl":124},"scripts\u002Flib\u002Felement-selector.mjs",2950,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Felement-selector.mjs",{"path":126,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":127,"sourceHtmlUrl":128},"scripts\u002Flib\u002Fevaluation.mjs",1972,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fevaluation.mjs",{"path":130,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":131,"sourceHtmlUrl":132},"scripts\u002Flib\u002Fextraction.mjs",1434,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fextraction.mjs",{"path":134,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":135,"sourceHtmlUrl":136},"scripts\u002Flib\u002Ffile-upload.mjs",1308,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Ffile-upload.mjs",{"path":138,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":139,"sourceHtmlUrl":140},"scripts\u002Flib\u002Fhtml-diff.mjs",3928,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fhtml-diff.mjs",{"path":142,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":143,"sourceHtmlUrl":144},"scripts\u002Flib\u002Fkey-definitions.mjs",3793,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fkey-definitions.mjs",{"path":146,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":147,"sourceHtmlUrl":148},"scripts\u002Flib\u002Fkeyboard-input.mjs",7191,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fkeyboard-input.mjs",{"path":150,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":151,"sourceHtmlUrl":152},"scripts\u002Flib\u002Fmouse.mjs",9592,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fmouse.mjs",{"path":154,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":155,"sourceHtmlUrl":156},"scripts\u002Flib\u002Fnavigation.mjs",6093,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fnavigation.mjs",{"path":158,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":159,"sourceHtmlUrl":160},"scripts\u002Flib\u002Fpage-scripts\u002Fdom-summary.mjs",1479,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fpage-scripts\u002Fdom-summary.mjs",{"path":162,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":163,"sourceHtmlUrl":164},"scripts\u002Flib\u002Fpage-scripts\u002Fmarkdown.mjs",2793,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fpage-scripts\u002Fmarkdown.mjs",{"path":166,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":167,"sourceHtmlUrl":168},"scripts\u002Flib\u002Fpage-scripts\u002Fpermission-shim.mjs",3102,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fpage-scripts\u002Fpermission-shim.mjs",{"path":170,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":171,"sourceHtmlUrl":172},"scripts\u002Flib\u002Fpage-session.mjs",2313,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fpage-session.mjs",{"path":174,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":175,"sourceHtmlUrl":176},"scripts\u002Flib\u002Fprofile-lock.mjs",2588,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fprofile-lock.mjs",{"path":178,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":179,"sourceHtmlUrl":180},"scripts\u002Flib\u002Fscreenshot.mjs",5250,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fscreenshot.mjs",{"path":182,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":183,"sourceHtmlUrl":184},"scripts\u002Flib\u002Fselect-option.mjs",2844,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fselect-option.mjs",{"path":186,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":187,"sourceHtmlUrl":188},"scripts\u002Flib\u002Fsession-state.mjs",792,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fsession-state.mjs",{"path":190,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":191,"sourceHtmlUrl":192},"scripts\u002Flib\u002Ftabs.mjs",3341,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Ftabs.mjs",{"path":194,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":195,"sourceHtmlUrl":196},"scripts\u002Flib\u002Fviewport.mjs",2824,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fviewport.mjs",{"path":198,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":199,"sourceHtmlUrl":200},"scripts\u002Flib\u002Fwebsocket-client.mjs",5129,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe-glass\u002F.cursor-plugin\u002Fskills\u002Fbrowsing\u002Fscripts\u002Flib\u002Fwebsocket-client.mjs",[202,209,216,223,230,237],{"skillNo":203,"name":204,"nameZh":6,"description":205,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":206,"categories":207,"keywords":208,"publishedAt":18},"01M7C1Z5362Y8NVRZ69HWP0ZD0","write-skill","Use when creating new skills, editing existing skills, or verifying skills work before deployment","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwrite-skill\u002FSKILL.md",[],[],{"skillNo":210,"name":211,"nameZh":6,"description":212,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":213,"categories":214,"keywords":215,"publishedAt":18},"01M7C1YZ9966WK9R6ZRPCGWF3G","write-plan","Use when you have a spec or requirements for a multi-step task, before touching code","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwrite-plan\u002FSKILL.md",[],[],{"skillNo":217,"name":218,"nameZh":6,"description":219,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":220,"categories":221,"keywords":222,"publishedAt":18},"01M7C1YMR4SJD7BPBBPXZCVGVH","write-clearly","Apply Strunk's timeless writing rules to ANY prose humans will read—documentation, commit messages, error messages, explanations, reports, or UI text. Makes your writing clearer, stronger, and more professional.","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwrite-clearly\u002FSKILL.md",[],[],{"skillNo":224,"name":225,"nameZh":6,"description":226,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":227,"categories":228,"keywords":229,"publishedAt":18},"01M7C1YH9805YSM0N20VQAVSXF","windows-vm","Create, manage, or connect to a headless Windows 11 VM running in Docker with SSH access. Use when the user wants to spin up, stop, restart, or SSH into a Windows VM.","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwindows-vm\u002FSKILL.md",[],[],{"skillNo":231,"name":232,"nameZh":6,"description":233,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":234,"categories":235,"keywords":236,"publishedAt":18},"01M7C1YD87SCNEJ2C9B8HFR2NQ","verify-completion","Use when about to claim work is complete, fixed, or passing, before committing or creating MRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fverify-completion\u002FSKILL.md",[],[],{"skillNo":238,"name":239,"nameZh":6,"description":240,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":241,"categories":242,"keywords":243,"publishedAt":18},"01M7C1YARDEHT0TJ9BRG30ZCF4","using-moe","Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fusing-moe\u002FSKILL.md",[],[]]