MCPInstall

vibe-check

Browser automation for AI agents. Use when the user needs to navigate websites, read page content, fill forms, click elements, take screenshots, or manage browser pages.

Installs
823
Last 24h
—
Trending rank
—
Security
Not audited

Install

Works with every agent the skills CLI supports.

npx skills add https://github.com/vibiumdev/vibium --skill vibe-check
  1. 01

    Install

    Run the command in your project. The CLI asks which agents to install for.

  2. 02

    Describe the task

    Ask your agent for something this skill covers. It loads vibe-check automatically when relevant.

  3. 03

    Review the result

    The agent follows the SKILL.md steps, scripts and references bundled with the skill.

SKILL.md

The instructions your agent reads when the skill is triggered.

Vibium Browser Automation — CLI Reference

The vibium CLI automates Chrome (and Firefox, via --engine firefox) from the command line. The browser auto-launches on first use (daemon mode keeps it running between commands).

vibium go <url> && vibium map && vibium click @e1 && vibium map

Core Workflow

Every browser automation follows this pattern:

  1. Navigate: vibium go <url>
  2. Map: vibium map (get element refs like @e1, @e2)
  3. Interact: Use refs to click, fill, select — e.g. vibium click @e1
  4. Re-map: After navigation or DOM changes, get fresh refs with vibium map

Binary Resolution

Before running any commands, resolve the vibium binary path once:

  1. Try vibium directly (works if globally installed via npm install -g vibium)
  2. Fall back to ./clicker/bin/vibium (dev environment, in project root)
  3. Fall back to ./node_modules/.bin/vibium (local npm install)

Run vibium --help (or the resolved path) to confirm. Use the resolved path for all subsequent commands.

Windows note: Use forward slashes in paths (e.g. ./clicker/bin/vibium.exe) and quote paths containing spaces.

Command Chaining

Chain commands with && to run them sequentially. The chain stops on first error:

vibium go https://example.com && vibium map && vibium click @e3 && vibium diff map

When to chain: Use && for sequences that should happen back-to-back (navigate → interact → verify). Run commands separately when you need to inspect output between steps.

When NOT to chain: Don't chain commands that depend on parsing the previous output (e.g. reading map output to decide what to click). Run those separately so you can analyze the result first.

Commands

Discovery

  • vibium map — map interactive elements with @refs (recommended before interacting)
  • vibium map --selector "nav" — scope map to elements within a CSS subtree
  • vibium diff map — compare current vs last map (see what changed)

Navigation

  • vibium go <url> — go to a page
  • vibium back — go back in history
  • vibium forward — go forward in history
  • vibium reload — reload the current page
  • vibium url — print current URL
  • vibium title — print page title

Reading Content

  • vibium text — get all page text
  • vibium text "<selector>" — get text of a specific element
  • vibium html — get page HTML (use --outer for outerHTML)
  • vibium find "<selector>" — find element, return @e1 ref (clickable with vibium click @e1)
  • vibium find "<selector>" --all — find all matching elements → @e1, @e2, ... (--limit N)
  • vibium find text "Sign In" — find element by text content → @e1
  • vibium find label "Email" — find input by label → @e1
  • vibium find placeholder "Search" — find by placeholder → @e1
  • vibium find testid "submit-btn" — find by data-testid → @e1
  • vibium find xpath "//div[@class]" — find by XPath → @e1
  • vibium find alt "Logo" — find by alt attribute → @e1
  • vibium find title "Settings" — find by title attribute → @e1
  • vibium find role <role> — find element by ARIA role → @e1 (--name for accessible name filter)
  • vibium eval "<js>" — run JavaScript and print result (--stdin to read from stdin)
  • vibium count "<selector>" — count matching elements
  • vibium screenshot -o file.png — capture screenshot (--full-page, --annotate)
  • vibium a11y-tree — accessibility tree (--everything for all nodes)

Interaction

  • vibium click "<selector>" — click an element (also accepts @ref from map)
  • vibium dblclick "<selector>" — double-click an element
  • vibium type "<selector>" "<text>" — type into an input (appends to existing value)
  • vibium fill "<selector>" "<text>" — clear field and type new text (replaces value)
  • vibium press <key> [selector] — press a key on element or focused element
  • vibium focus "<selector>" — focus an element
  • vibium hover "<selector>" — hover over an element
  • vibium scroll [direction] — scroll page (--amount N, --selector)
  • vibium scroll into-view "<selector>" — scroll element into view (centered)
  • vibium keys "<combo>" — press keys (Enter, Control+a, Shift+Tab)
  • vibium select "<selector>" "<value>" — pick a dropdown option
  • vibium check "<selector>" — check a checkbox/radio (idempotent)
  • vibium uncheck "<selector>" — uncheck a checkbox (idempotent)

Mouse Primitives

  • vibium mouse click [x] [y] — click at coordinates or current position (--button 0|1|2)
  • vibium mouse move <x> <y> — move mouse to coordinates
  • vibium mouse down — press mouse button (--button 0|1|2)
  • vibium mouse up — release mouse button (--button 0|1|2)
  • vibium drag "<source>" "<target>" — drag from one element to another

Element State

  • vibium value "<selector>" — get input/textarea/select value
  • vibium attr "<selector>" "<attribute>" — get HTML attribute value
  • vibium is visible "<selector>" — check if element is visible (true/false)
  • vibium is enabled "<selector>" — check if element is enabled (true/false)
  • vibium is checked "<selector>" — check if checkbox/radio is checked (true/false)
  • vibium is actionable "<selector>" — check if element is actionable (true/false)

Waiting

  • vibium wait "<selector>" — wait for element (--state visible|hidden|attached, --timeout ms)
  • vibium wait url "<pattern>" — wait until URL contains substring (--timeout ms)
  • vibium wait load — wait until page is fully loaded (--timeout ms)
  • vibium wait text "<text>" — wait until text appears on page (--timeout ms)
  • vibium wait fn "<expression>" — wait until JS expression returns truthy (--timeout ms)
  • vibium sleep <ms> — pause execution (max 30000ms)

Capture

  • vibium screenshot -o file.png — capture screenshot (--full-page, --annotate)
  • vibium pdf -o file.pdf — save page as PDF

Dialogs

  • vibium dialog accept [text] — accept dialog (optionally with prompt text)
  • vibium dialog dismiss — dismiss dialog

Emulation

  • vibium viewport — get current viewport dimensions
  • vibium viewport <width> <height> — set viewport size (--dpr for device pixel ratio)
  • vibium window — get OS browser window dimensions and state
  • vibium window <width> <height> [x] [y] — set window size and position (--state)
  • vibium media — override CSS media features (--color-scheme, --reduced-motion, --forced-colors, --contrast, --media)
  • vibium geolocation <lat> <lng> — override geolocation (--accuracy)
  • vibium content "<html>" — replace page HTML (--stdin to read from stdin)

Frames

  • vibium frames — list all iframes on the page
  • vibium frame "<nameOrUrl>" — find a frame by name or URL substring

File Upload

  • vibium upload "<selector>" <files...> — set files on input[type=file]

Recording

  • vibium record start — start recording (--screenshots, --snapshots, --name, -o path — defaults to a timestamped record-<date>.zip, so reruns never overwrite)
  • vibium record stop — stop recording and save ZIP (-o path overrides the start path; the output names the saved file)

Recordings can include a video track of the session (Firefox 154+, local browsers). By default video is recorded when the engine supports it and skipped otherwise — the stop result says which. Pass --video to require it (fails with an explanatory error on Chrome), --video=false to disable, and --video-size 1280x720 / --video-fps 30 to override the viewport defaults. The video lands inside the recording ZIP next to the trace (video/<context>.webm); it films the page that was active at start and does not follow tab switches. Remote browser connections record every track except video; --video-remote keep records anyway and leaves the file on the remote host (the stop output names its path there).

vibium record start --video -o run.zip
# ... actions ...
vibium record stop
# Saved run.zip (23 steps, 14s video)

Cookies

  • vibium cookies — list all cookies
  • vibium cookies <name> <value> — set a cookie
  • vibium cookies clear — clear all cookies

Storage State

  • vibium storage — export cookies + localStorage + sessionStorage (-o state.json)
  • vibium storage restore <path> — restore state from JSON file

Downloads

  • vibium download dir <path> — set download directory

Pages

  • vibium pages — list open pages
  • vibium page new [url] — open new page
  • vibium page new --isolated [url] — open page with its own cookies/storage
  • vibium page switch <index|url> — switch page
  • vibium page close [index|page id] — close page

Debug

  • vibium highlight "<selector>" — highlight element visually (3 seconds)

Session

  • vibium start — start a local browser session
  • vibium start <url> — start connected to a remote browser
  • vibium stop — stop the browser session
  • vibium daemon start — start background browser
  • vibium daemon status — check if running
  • vibium daemon stop — stop daemon
  • vibium --session <name> <command> — run against an isolated daemon and browser

Common Patterns

Ref-based workflow (recommended for AI)

vibium go https://example.com
vibium map
vibium click @e1
vibium map  # re-map after interaction

Verify action worked

vibium map
vibium click @e3
vibium diff map  # see what changed

Read a page

vibium go https://example.com && vibium text

Fill a form (end-to-end)

vibium go https://example.com/login
vibium map
# Look at map output to identify form fields
vibium fill @e1 "user@example.com"
vibium fill @e2 "secret"
vibium click @e3
vibium wait url "/dashboard"
vibium screenshot -o after-login.png

Scoped map (large pages)

vibium map --selector "nav"        # Only map elements in <nav>
vibium map --selector "#sidebar"   # Only map elements in #sidebar
vibium map --selector "form"       # Only map form controls

Semantic find (no CSS selectors needed)

vibium find text "Sign In"             # → @e1 [button] "Sign In"
vibium find label "Email"              # → @e1 [input] placeholder="Email"
vibium click @e1                       # Click the found element
vibium find placeholder "Search..."    # → @e1 [input] placeholder="Search..."
vibium find testid "submit-btn"        # → @e1 [button] "Submit"
vibium find alt "Company logo"         # → @e1 [img] alt="Company logo"
vibium find title "Close"              # → @e1 [button] title="Close"
vibium find xpath "//a[@href='/about']"  # → @e1 [a] "About"

Authentication with state persistence

# Log in once and save state
vibium go https://app.example.com/login
vibium fill "input[name=email]" "user@example.com"
vibium fill "input[name=password]" "secret"
vibium click "button[type=submit]"
vibium wait url "/dashboard"
vibium storage -o auth.json

# Restore in a later session (skips login)
vibium storage restore auth.json
vibium go https://app.example.com/dashboard

Extract structured data

vibium go https://example.com
vibium eval "JSON.stringify([...document.querySelectorAll('a')].map(a => ({text: a.textContent.trim(), href: a.href})))"

Check page structure without rendering

vibium go https://example.com && vibium a11y-tree

Remote browser

vibium start ws://remote-host:9515/session
vibium go https://example.com
vibium map
vibium stop

Concurrent sessions (isolated browsers)

# Two scripts on one host, each with its own daemon and browser
export VIBIUM_SESSION=checkout-tests
vibium go https://example.com/checkout
vibium daemon stop

# Or per command, to drive two browsers from one script
vibium --session buyer go https://shop.example.com
vibium --session seller go https://shop.example.com/admin

Isolated pages (one browser, separate cookies/storage)

# Lighter than a session: two logins in one browser, no second launch
vibium page new --isolated https://shop.example.com   # prints (page: <id>)
vibium page new --isolated https://shop.example.com
vibium page close <id>   # also removes the page's isolated context

Multi-page workflow

vibium page new https://docs.example.com
vibium text "h1"
vibium page switch 0

Annotated screenshot

vibium screenshot -o annotated.png --annotate

Inspect an element

vibium attr "a" "href"
vibium value "input[name=email]"
vibium is visible ".modal"

Save as PDF

vibium go https://example.com && vibium pdf -o page.pdf

Eval / JavaScript

vibium eval is the escape hatch for any DOM query or mutation the CLI doesn't cover directly.

Simple expressions — use single quotes:

vibium eval 'document.title'
vibium eval 'document.querySelectorAll("li").length'

Complex scripts — use --stdin with a heredoc:

vibium eval --stdin <<'EOF'
const rows = [...document.querySelectorAll('table tbody tr')];
JSON.stringify(rows.map(r => {
  const cells = r.querySelectorAll('td');
  return { name: cells[0].textContent.trim(), price: cells[1].textContent.trim() };
}));
EOF

JSON output — use --json to get machine-readable output:

vibium eval --json 'JSON.stringify({url: location.href, title: document.title})'

Important: eval returns the expression result. If your script doesn't return a value, you'll get null. Always make sure the last expression evaluates to the data you want.

Timeouts and Waiting

All interaction commands (click, fill, type, etc.) auto-wait for the target element to be actionable. You usually don't need explicit waits.

Use explicit waits when:

  • Waiting for navigation: vibium wait url "/dashboard" — after clicking a link that navigates
  • Waiting for content: vibium wait text "Success" — after form submission, wait for confirmation
  • Waiting for element: vibium wait ".modal" — wait for a modal to appear
  • Waiting for page load: vibium wait load — after navigation to a slow page
  • Waiting for JS condition: vibium wait fn "window.appReady === true" — wait for app initialization
  • Fixed delay (last resort): vibium sleep 2000 — only when no better signal exists (max 30s)

All wait commands accept --timeout <ms> (default varies by command).

Ref Lifecycle

Refs (@e1, @e2) are invalidated when the page changes. Always re-map after:

  • Clicking links or buttons that navigate
  • Form submissions
  • Dynamic content loading (dropdowns, modals)

Global Flags

FlagDescription
--engine <name>Browser engine: chrome (default) or firefox (env: VIBIUM_ENGINE)
--channel <ch>Engine release channel — Firefox: release (default) or beta; Chrome: stable (default), beta, dev, or canary (env: VIBIUM_ENGINE_CHANNEL)
--headlessHide browser window
--jsonOutput as JSON
-v, --verboseDebug logging
--session <name>Isolated daemon + browser for concurrent use (env: VIBIUM_SESSION)

Tips

  • All click/type/hover/fill actions auto-wait for the element to be actionable
  • All selector arguments also accept @ref from vibium map
  • Use vibium map before interacting to discover interactive elements
  • Use vibium map --selector to reduce noise on large pages
  • Use vibium fill to replace a field's value, vibium type to append to it
  • Use vibium find text / find label / find testid for semantic element lookup (more reliable than CSS selectors)
  • Use vibium find role for ARIA-role-based lookup
  • Use vibium a11y-tree to understand page structure without visual rendering
  • Use vibium text "<selector>" to read specific sections
  • Use vibium diff map after interactions to see what changed
  • vibium eval is the escape hatch for complex DOM queries
  • vibium check/vibium uncheck are idempotent — safe to call without checking state first
  • Screenshots save to the current directory by default (-o to change)
  • Use vibium storage / vibium storage restore to persist auth across sessions
  • Without --session, all commands on a host share one daemon and one browser — set VIBIUM_SESSION when running concurrently

FAQ

What does the vibe-check skill do?

Browser automation for AI agents. Use when the user needs to navigate websites, read page content, fill forms, click elements, take screenshots, or manage browser pages.

How do I install vibe-check?

Run npx skills add https://github.com/vibiumdev/vibium --skill vibe-check in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/vibe-check/ (personal) or .claude/skills/vibe-check/ (project).

Which AI agents can use vibe-check?

Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.

Is vibe-check safe to install?

No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.

Who maintains vibe-check?

It is published by vibiumdev in vibiumdev/vibium. It has 823 recorded installs.

Popular productivity & automation skills

View category →

agent-browser

vercel-labs/agent-browser

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.

OfficialProductivity & Automation1.3M installs

lark-doc

open.feishu.cn

飞书云文档(Docx / Wiki)内容操作:读取、创建、编辑文档,插入或下载图片附件,以及操作思维笔记。用户提供文档 URL/token(包括 doubao.com 的 /docx/、/wiki/)时使用;按 URL 路径/token 而非域名路由;/page/ 是妙搭应用,非飞书云文档。文档内嵌资源按读取参考中的统一规则分流。独立评论操作走 lark-drive;随正文读取评论使用 docs +fetch。表格或 Base 内部数据操作不在本 skill。

lark-base

open.feishu.cn

飞书多维表格(Base)操作:建表、字段、记录、视图、统计、公式/lookup、表单、仪表盘、应用模式(BaseApp/AppMode 页面与组件)、Workspace 目录、workflow、角色权限、模板中心(多维表格模板分类/列表/搜索);遇到 Base/多维表格/bitable、BaseApp/AppMode、/base/ 或 /app/ 链接时使用。BaseApp 不走 lark-apps;文件导入/导出转 lark-drive,认证/授权转 lark-shared。

lark-drive

open.feishu.cn

飞书云空间(云盘/云存储):管理 Drive 文件和文件夹,包含上传/下载、创建文件夹、复制/移动/删除、查看元数据、查询权限设置、评论/权限/订阅、标题、版本、飞书文档密级标签(secure labels)和本地文件导入。用户需要整理云盘目录、处理云空间资源 URL/token、判断链接类型/真实 token/标题,或导入 Word/Markdown/Excel/CSV/PPTX/.base 为 docx/sheet/bitable/slides 时使用;doubao.com 云空间 URL/token 也按资源路径和 token 路由,不回退 WebFetch。不负责:文档内容编辑(走 lark-doc)、表格/Base 表内数据操作(走 lark-sheets/lark-base)、知识空间节点/成员管理(走 lark-wiki)、原生 Markdown 文件读写/patch/diff(走 lark-markdown)。

lark-shared

open.feishu.cn

Use for lark-cli setup/auth tasks: auth login/status/logout, user vs bot identity, business-domain permissions (--domain, including all/docs/drive), missing scopes, revoking authorization, or handling notice JSON.

lark-im

open.feishu.cn

飞书即时通讯:收发消息和管理群聊。发送和回复消息、搜索聊天记录、管理群聊成员、上传下载图片和文件、管理表情回复、发送应用内/短信/电话加急、发送和处理交互卡片(Interactive Card)、监听卡片按钮回调(card.action.trigger)。当用户需要发消息、查看或搜索聊天记录、下载聊天中的文件、查看群成员、搜索群、创建群聊或话题群、管理标记数据、管理 Feed 置顶(添加/移除/查询置顶会话)、管理标签数据、处理卡片回调时使用。