Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
runablehq avatar

Mini Browser

  • 2.4k installs
  • 182 repo stars
  • Updated May 30, 2026
  • runablehq/mini-browser

mini-browser is an agent browser automation skill using the mb CLI over Chrome CDP for navigation, forms, and audits.

About

mini-browser provides the mb CLI for agent browser automation over Chrome DevTools Protocol on port 9222 via puppeteer-core. Setup checks which mb and curl to 127.0.0.1:9222/json/version, installing @runablehq/mini-browser and running mb-start-chrome when missing. Navigation commands include mb go, url, back, and forward. Observation uses mb text, mb shot screenshots, and mb snap listing interactive elements with coordinates. Interaction covers click, type, fill key=value fields, key presses, move, drag, and scroll. Recording supports start and stop screencasts to webm, mp4, or gif. mb audit runs design checks for palette, typography, contrast, accessibility, and SEO with optional JSON output. The observe-act loop snapshots, clicks coordinates, waits networkidle, and repeats. Viewport is 1024 by 768 so snap only lists visible elements until scroll. fill matches aria-label, placeholder, name, id, label text, or CSS selectors.

  • mb CLI Unix-style commands over CDP Chrome on 9222.
  • Observe-act loop with snap coordinates and click.
  • mb fill matches labels placeholders and aria attributes.
  • mb audit for contrast accessibility and SEO checks.
  • Screencast record start stop with fps and scale flags.

Mini Browser by the numbers

  • 2,430 all-time installs (skills.sh)
  • Ranked #170 of 2,715 Automation & Workflows skills by installs in the Skillselion catalog
  • Security screen: HIGH risk (skills.sh audit)
  • Data as of Aug 4, 2026 (Skillselion catalog sync)
At a glance

mini-browser capabilities & compatibility

Capabilities
cdp chrome launch and health checks · navigation and url control commands · coordinate based click type and fill interaction · screenshot and visible text extraction · design and accessibility audit reports · screencast recording with fps and scale options
Works with
chrome · playwright
Use cases
web scraping · testing · ui design
Pricing
Free
From the docs

What mini-browser says it does

Viewport is 1024×768.
SKILL.md
mb audit # human-readable report
SKILL.md
npx skills add https://github.com/runablehq/mini-browser --skill mini-browser

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs2.4k
repo stars182
Security audit1 / 3 scanners passed
Last updatedMay 30, 2026
Repositoryrunablehq/mini-browser

How do I let an agent browse, screenshot, fill forms, and audit pages from the terminal?

Automate browsing with the mb CLI via CDP for navigation, screenshots, forms, audits, and screencasts.

Who is it for?

Agents and developers needing lightweight terminal browser automation without full Playwright scripts.

Skip if: Skip for non-Chrome browsers, mobile native app UI, or production load testing at scale.

When should I use this skill?

User asks to browse, screenshot, scrape, fill forms, record screen, or run design audit with mb.

What you get

Chrome session controlled via mb with navigation, interaction, capture, or audit output as requested.

  • page screenshots
  • scraped text
  • screencast recordings

Files

SKILL.mdMarkdownGitHub ↗

mini-browser (mb) — Browser CLI for Agents

mb is a browser CLI where each command is a small Unix tool. It talks to Chrome over CDP (port 9222) via puppeteer-core.

Setup (only if not already available)

Setup is only needed when mb is not installed or Chrome is not reachable. Run these checks first — if both pass, skip straight to the Command Reference.

Check if ready

# 1. Is mb installed?
which mb && echo "mb: ok" || echo "mb: MISSING"

# 2. Is Chrome listening on CDP?
curl -sf http://127.0.0.1:9222/json/version > /dev/null && echo "chrome: ok" || echo "chrome: NOT RUNNING"

If both print "ok", everything is ready — go use mb commands directly.

Install (only if mb is missing)

npm install -g @runablehq/mini-browser

Start Chrome (only if not running)

mb-start-chrome

This launches Chrome with --remote-debugging-port=9222, a fresh profile, and a 1024×768 window. It no-ops if Chrome is already running.

To kill and relaunch:

mb-restart-chrome

Verify

mb go "https://example.com" && mb text

Environment Variables

VariableDefaultDescription
CHROME_PORT9222CDP port
CHROME_BINauto-detectedPath to Chrome/Chromium binary
CHROME_PID_FILE<scripts>/.chrome-pidPID file location
CHROME_USER_DATA_DIR<scripts>/.chrome-profileChrome profile directory

Command Reference

Navigation

CommandDescription
mb go <url>Navigate to URL (waits for networkidle)
mb urlPrint current URL
mb backGo back
mb forwardGo forward

Observation

CommandDescription
mb text [selector]Visible text content (default: body)
mb shot [file]Screenshot to PNG (default: ./shot.png)
mb snapList interactive elements with coordinates

Interaction

CommandDescription
mb click <x> <y>Click at coordinates
mb type [x y] <text>Type text (with coords: selects first)
mb fill <k=v...>Fill form fields by label/name/placeholder
mb key <key...>Press keys (Enter, Tab, Meta+a)
mb move <x> <y>Hover at coordinates
mb drag <x1> <y1> <x2> <y2>Drag between points
mb scroll [dir] [px]Scroll (default: down 500)

Recording

CommandDescription
mb record start <file>Start recording (.webm, .mp4, .gif)
mb record stopStop recording and save
mb record statusCheck if recording is active

Tabs

CommandDescription
mb tab listList open tabs
mb tab new [url]Open new tab, print index
mb tab close [n]Close tab (default: last)

Other

CommandDescription
mb js <code>Run JavaScript in page context
mb wait <target>Wait for ms / selector / networkidle / url:pattern
mb auditDesign audit (palette, typography, contrast, a11y, SEO)
mb logsStream console logs (Ctrl+C to stop)

Flags

FlagDefaultDescription
--timeout <ms>30000Command timeout
--tab <n>0Target tab index
--jsonfalseStructured JSON output
--rightfalseRight-click
--doublefalseDouble-click
--fps <n>30Recording frame rate
--scale <n>1Recording scale factor

Usage Patterns

Observe → Act loop

The standard agent loop: snapshot the page, pick an element, act on it.

mb snap                          # list interactive elements with (x, y)
mb click 512 380                 # click the button at those coordinates
mb wait networkidle              # wait for the page to settle
mb snap                          # observe again

Fill and submit a form

mb go "https://example.com/login"
mb fill "Email=user@example.com" "Password=hunter2"
mb key Enter
mb wait url:/dashboard

Take a screenshot

mb shot page.png
mb shot page.png --width 1440 --height 900

Extract text

mb text "main"                   # text from <main>
mb text "#content"               # text from #content
mb text                          # full body text

Run JavaScript

mb js 'document.title'
echo 'document.querySelectorAll("a").length' | mb js -

Record a screencast

mb record start demo.mp4 --fps 30 --scale 1
# ... interact with the page ...
mb record stop

Design audit

mb audit                         # human-readable report
mb audit --json                  # structured JSON output

Dismiss overlays

Cookie banners and modals block clicks. Remove them with JS:

mb js 'document.querySelector("[class*=cookie]")?.remove()'

Wait strategies

mb wait 2000                     # sleep 2 seconds
mb wait ".modal"                 # wait for selector to appear
mb wait networkidle              # wait for no network activity
mb wait url:/dashboard           # wait for URL to contain string

Important Notes

  • Viewport is 1024×768. snap only returns elements in the current viewport — scroll and snap again to find more.
  • `text` uses querySelector — returns first match only. Use text "main" over text "p" for better results.
  • `go` waits for networkidle. For heavy SPAs, follow up with wait ".selector".
  • `type` with coordinates triple-clicks first to select existing text, then types the replacement.
  • `fill` field matching order: aria-label → placeholder → name attr → id → label text → CSS selector (use #/./[ prefix).
  • `--json` output: snap[{role, name, x, y, state}], tab list[{index, url, title}], logs → JSON lines, audit → full audit object.
  • Recording state is stored in ~/.mb-recorder.json. Only one recording at a time.
  • `tab close` cannot close the last remaining tab.

Troubleshooting

ProblemFix
"Chrome not found"Set CHROME_BIN=/path/to/chrome
Connection refusedRun mb-start-chrome first
Stale recording stateDelete ~/.mb-recorder.json
Chrome window wrong sizemb-restart-chrome (creates fresh profile)
Element not in snap outputmb scroll down 500 then mb snap again

Related skills

How it compares

Pick mini-browser for agent-operated terminal browser tasks; choose Playwright skills when you need full cross-browser test frameworks in CI.

FAQ

How verify mb is ready?

which mb succeeds and curl to 127.0.0.1:9222/json/version returns ok.

Why scroll before snap again?

Viewport is 1024x768; snap only lists elements currently visible.

How dismiss cookie banners blocking clicks?

Use mb js to remove overlay nodes before clicking targets.

Is Mini Browser safe to install?

skills.sh reports 1 of 3 security scanners passed. Review the Security Audits panel on this page before installing in production.

Automation & Workflowsintegrationstesting

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.