MCPInstall

chrome-devtools-cli

by reason-machinesProductivity & Automation

Chrome DevTools Protocol CLI that auto-connects to existing Chrome for browser automation without heavyweight runtime overhead

Installs
207
Last 24h
—
Trending rank
—
Security
Not audited

Install

Works with every agent the skills CLI supports.

npx skills add https://github.com/reason-machines/devtools-skills --skill chrome-devtools-cli
  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 chrome-devtools-cli 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.

Chrome DevTools CLI

Skill by ara.so — Devtools Skills collection.

Chrome DevTools CLI is a lightweight Rust binary that connects directly to an existing Chrome browser via the DevTools Protocol. It auto-connects to your Chrome instance, requires no separate browser process, and provides a minimal-context alternative to MCP-based browser automation.

Why Use This

  • Zero context overhead: One command in, one result out. No MCP layer, no large protocol payloads
  • Uses your Chrome: Works with your existing browser session, credentials, and extensions
  • Persistent connection: Background daemon keeps WebSocket open, browser prompts for access only once
  • Deterministic targeting: Friendly target names ([target:red-snake]) for reliable multi-page workflows

Installation

# macOS (recommended)
brew install aeroxy/tap/chrome-devtools

# Or via Cargo
cargo install chrome-devtools-cli

The binary is named chrome-devtools.

Prerequisites

Enable Chrome remote debugging:

  1. Open Chrome
  2. Navigate to chrome://inspect/#remote-debugging
  3. Enable the remote debugging server

Auto-Connect Behavior

The CLI automatically discovers Chrome's WebSocket endpoint by reading DevToolsActivePort from:

  • macOS: ~/Library/Application Support/Google/Chrome/
  • Linux: ~/.config/google-chrome/
  • Windows: %LOCALAPPDATA%\Google\Chrome\User Data\

Override with environment variables or flags:

# Environment variables
export CHROME_WS_ENDPOINT=ws://localhost:9222/devtools/browser/abc123
export CHROME_USER_DATA_DIR=~/custom-chrome-profile
export CHROME_CHANNEL=beta  # stable, beta, canary, dev

# Or flags
chrome-devtools --ws-endpoint ws://localhost:9222/... navigate https://example.com
chrome-devtools --user-data-dir ~/custom-profile navigate https://example.com
chrome-devtools --channel canary navigate https://example.com

Daemon Architecture

First command spawns a background daemon that:

  • Connects to Chrome WebSocket (one-time approval prompt)
  • Listens on Unix socket: /tmp/chrome-devtools-daemon.sock
  • Auto-exits after 5 minutes of inactivity
  • All subsequent commands reuse the persistent connection

Kill manually: pkill -f __daemon__ or delete the socket file.

Page Targeting Pattern

CRITICAL: Always capture and reuse the [target:name] from navigation commands.

# Step 1: Navigate and capture target name
chrome-devtools navigate https://github.com
# Output: Navigated to https://github.com
# [target:green-dog]

# Step 2: Pin all subsequent commands to this target
chrome-devtools --target green-dog screenshot --output /tmp/gh.png
chrome-devtools --target green-dog evaluate "document.title"
chrome-devtools --target green-dog click "a[href='/login']"

Without --target, commands default to page index 0, which may change as Chrome reorders tabs.

List All Pages

chrome-devtools list-pages
# Output:
# [0] (green-dog) GitHub — https://github.com
# [1] (red-snake) Example Domain — https://example.com
# [2] (bold-stag) Local Dev — https://localhost:3000

You can also use --page <index> for quick one-offs:

chrome-devtools --page 1 evaluate "window.location.href"

Core Commands

Navigation

# Navigate to URL (waits for load event)
chrome-devtools navigate https://example.com

# History navigation
chrome-devtools --target red-snake navigate --back
chrome-devtools --target red-snake navigate --forward
chrome-devtools --target red-snake navigate --reload

# Tab management
chrome-devtools new-page https://github.com
chrome-devtools close-page 2
chrome-devtools select-page 0  # Bring tab to front

Inspection

# Screenshot (viewport only)
chrome-devtools --target green-dog screenshot --output /tmp/page.png

# Full-page screenshot (scrollable area)
chrome-devtools --target green-dog screenshot --full-page --output /tmp/full.png

# JavaScript evaluation
chrome-devtools --target green-dog evaluate "document.title"
chrome-devtools --target green-dog evaluate "Array.from(document.querySelectorAll('h1')).map(h => h.textContent)"

# Handle JavaScript dialogs (alert/confirm/prompt)
chrome-devtools --target green-dog evaluate "confirm('Proceed?')" --dialog-action accept
chrome-devtools --target green-dog evaluate "prompt('Name?')" --dialog-action "John Doe"

# Accessibility tree snapshot
chrome-devtools --target green-dog snapshot

Interaction

# Click by CSS selector
chrome-devtools --target green-dog click "#submit-button"
chrome-devtools --target green-dog click "a[href='/login']"

# Click at coordinates
chrome-devtools --target green-dog click-at 100 200

# Fill input fields
chrome-devtools --target green-dog fill "#email" "user@example.com"
chrome-devtools --target green-dog fill "#password" "secret123"

# Fill dropdown (select element)
chrome-devtools --target green-dog fill "#country" "United States"

# Toggle checkbox/radio
chrome-devtools --target green-dog fill "#terms" "true"
chrome-devtools --target green-dog fill "#newsletter" "false"

# Type text into focused element
chrome-devtools --target green-dog type-text "Hello World"
chrome-devtools --target green-dog type-text "Search query" --submit-key Enter

# Press keys
chrome-devtools --target green-dog press-key Enter
chrome-devtools --target green-dog press-key "Control+A"
chrome-devtools --target green-dog press-key Escape

# Hover over element
chrome-devtools --target green-dog hover ".dropdown-menu"

Viewport & Waiting

# Resize viewport
chrome-devtools --target green-dog resize 1920 1080

# Wait for text to appear (default 30s timeout)
chrome-devtools --target green-dog wait-for "Login successful"
chrome-devtools --target green-dog wait-for "Dashboard" --timeout 60000

Third-Party Developer Tools

For pages that inject custom DevTools via window.__dtmcp:

# List available custom tools
chrome-devtools --target green-dog list-3p-tools

# Execute custom tool
chrome-devtools --target green-dog execute-3p-tool "myTool" '{"param": "value"}'

JSON Output Mode

Add --json to any command for machine-readable output:

chrome-devtools --json --target green-dog evaluate "document.title"
# {"success": true, "result": "Example Domain"}

chrome-devtools --json list-pages
# {"success": true, "pages": [{"index": 0, "friendlyName": "green-dog", ...}]}

Real-World Workflows

Automated Form Submission

#!/bin/bash
set -e

# Navigate and capture target
TARGET=$(chrome-devtools navigate https://example.com/form | grep -oP '(?<=target:)[a-z-]+')

# Fill form fields
chrome-devtools --target "$TARGET" fill "#name" "Alice Smith"
chrome-devtools --target "$TARGET" fill "#email" "alice@example.com"
chrome-devtools --target "$TARGET" fill "#country" "Canada"
chrome-devtools --target "$TARGET" fill "#terms" "true"

# Submit and wait for confirmation
chrome-devtools --target "$TARGET" click "#submit"
chrome-devtools --target "$TARGET" wait-for "Thank you" --timeout 10000

# Capture confirmation screenshot
chrome-devtools --target "$TARGET" screenshot --output /tmp/confirmation.png

Data Extraction

#!/bin/bash
TARGET=$(chrome-devtools navigate https://news.ycombinator.com | grep -oP '(?<=target:)[a-z-]+')

# Extract top stories
chrome-devtools --target "$TARGET" evaluate "
  Array.from(document.querySelectorAll('.athing')).slice(0, 5).map(item => ({
    title: item.querySelector('.titleline > a').textContent,
    url: item.querySelector('.titleline > a').href
  }))
" --json > /tmp/hn-stories.json

cat /tmp/hn-stories.json

Multi-Page Session

#!/bin/bash

# Open multiple pages
TARGET_HOME=$(chrome-devtools navigate https://github.com | grep -oP '(?<=target:)[a-z-]+')
TARGET_REPO=$(chrome-devtools new-page https://github.com/torvalds/linux | grep -oP '(?<=target:)[a-z-]+')

# Interact with different pages
chrome-devtools --target "$TARGET_HOME" click "a[href='/login']"
chrome-devtools --target "$TARGET_REPO" evaluate "document.querySelector('.repository-content').textContent.includes('Linux kernel')"

# Screenshot both
chrome-devtools --target "$TARGET_HOME" screenshot --output /tmp/home.png
chrome-devtools --target "$TARGET_REPO" screenshot --output /tmp/repo.png

Accessibility Audit

#!/bin/bash
TARGET=$(chrome-devtools navigate https://example.com | grep -oP '(?<=target:)[a-z-]+')

# Get full accessibility tree
chrome-devtools --target "$TARGET" snapshot > /tmp/a11y-tree.txt

# Check for specific roles
chrome-devtools --target "$TARGET" evaluate "
  document.querySelectorAll('[role=\"button\"]').length
"

# Find missing alt text
chrome-devtools --target "$TARGET" evaluate "
  Array.from(document.querySelectorAll('img:not([alt])')).map(img => img.src)
"

Common Patterns for AI Agents

1. Always Capture Target Names

# ❌ BAD: Target may change
chrome-devtools navigate https://example.com
chrome-devtools screenshot --output /tmp/page.png  # Might screenshot wrong page!

# ✅ GOOD: Explicit target
TARGET=$(chrome-devtools navigate https://example.com | grep -oP '(?<=target:)[a-z-]+')
chrome-devtools --target "$TARGET" screenshot --output /tmp/page.png

2. Understand the Page First

# Get context before interacting
chrome-devtools --target "$TARGET" snapshot  # See accessible structure
chrome-devtools --target "$TARGET" screenshot --output /tmp/context.png
chrome-devtools --target "$TARGET" evaluate "document.body.innerText"  # Read visible text

3. Wait for Dynamic Content

# Don't assume instant rendering
chrome-devtools --target "$TARGET" click "#load-more"
chrome-devtools --target "$TARGET" wait-for "Showing 20 results" --timeout 5000
chrome-devtools --target "$TARGET" evaluate "document.querySelectorAll('.result-item').length"

4. Handle Errors Gracefully

# Check page state before interaction
HAS_BUTTON=$(chrome-devtools --target "$TARGET" evaluate "!!document.querySelector('#submit')" --json | jq -r '.result')

if [ "$HAS_BUTTON" = "true" ]; then
  chrome-devtools --target "$TARGET" click "#submit"
else
  echo "Submit button not found"
fi

Troubleshooting

"Could not connect to Chrome"

  • Ensure Chrome is running
  • Verify remote debugging is enabled at chrome://inspect/#remote-debugging
  • Check DevToolsActivePort exists in user data directory
  • Try explicit WebSocket: chrome-devtools --ws-endpoint ws://localhost:9222/... navigate https://example.com

"Daemon connection failed"

# Kill stale daemon
pkill -f __daemon__
rm /tmp/chrome-devtools-daemon.sock

# Retry command (will spawn fresh daemon)
chrome-devtools navigate https://example.com

"Element not found" on click/fill

# Verify element exists
chrome-devtools --target "$TARGET" evaluate "!!document.querySelector('#my-button')"

# Get all matching elements
chrome-devtools --target "$TARGET" evaluate "document.querySelectorAll('#my-button').length"

# Wait for element to appear
chrome-devtools --target "$TARGET" wait-for "Submit" --timeout 10000
chrome-devtools --target "$TARGET" click "#submit"

Target name confusion

# List all pages with friendly names
chrome-devtools list-pages

# Use explicit page index if needed
chrome-devtools --page 0 evaluate "document.title"

# Or use raw target ID (from list-pages output)
chrome-devtools --target "E4F5A6B7C8D9" evaluate "document.title"

Best Practices for AI Agents

  1. Always use --target after initial navigation
  2. Capture screenshots before and after interactions for debugging
  3. Use snapshot to understand page structure before clicking
  4. Wait for content instead of assuming instant load
  5. Verify elements exist before interaction
  6. Use JSON output (--json) for programmatic parsing
  7. Check list-pages when multi-page workflows get confusing
  8. Kill daemon between unrelated sessions to avoid stale connections

Integration with AI Coding Agents

When helping a user automate Chrome:

  1. Install chrome-devtools-cli if not present
  2. Verify Chrome remote debugging is enabled
  3. Start with navigate and capture the target name
  4. Use snapshot or screenshot to understand the page
  5. Perform interactions with explicit --target
  6. Extract results with evaluate and --json
  7. Clean up with close-page if needed

Example agent workflow:

# User: "Login to example.com with my credentials"

# 1. Navigate
TARGET=$(chrome-devtools navigate https://example.com/login | grep -oP '(?<=target:)[a-z-]+')

# 2. Understand page
chrome-devtools --target "$TARGET" snapshot | head -50

# 3. Fill credentials (from env vars)
chrome-devtools --target "$TARGET" fill "#username" "$EXAMPLE_USERNAME"
chrome-devtools --target "$TARGET" fill "#password" "$EXAMPLE_PASSWORD"

# 4. Submit
chrome-devtools --target "$TARGET" click "#login-button"

# 5. Verify success
chrome-devtools --target "$TARGET" wait-for "Dashboard" --timeout 10000
chrome-devtools --target "$TARGET" screenshot --output /tmp/logged-in.png

This skill enables zero-context browser automation directly from your existing Chrome session.

FAQ

What does the chrome-devtools-cli skill do?

Chrome DevTools Protocol CLI that auto-connects to existing Chrome for browser automation without heavyweight runtime overhead

How do I install chrome-devtools-cli?

Run npx skills add https://github.com/reason-machines/devtools-skills --skill chrome-devtools-cli 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/chrome-devtools-cli/ (personal) or .claude/skills/chrome-devtools-cli/ (project).

Which AI agents can use chrome-devtools-cli?

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 chrome-devtools-cli 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 chrome-devtools-cli?

It is published by reason-machines in reason-machines/devtools-skills. It has 207 recorded installs.

More from reason-machines/devtools-skills

View all →

cc-switch-cli

reason-machines/devtools-skills

Open the skill to read its SKILL.md instructions.

315 installs

wx-cli-wechat-local-data

reason-machines/devtools-skills

Open the skill to read its SKILL.md instructions.

284 installs

chrome-devtools-mcp-automation

reason-machines/devtools-skills

Open the skill to read its SKILL.md instructions.

AI & Agents263 installs

officecli-office-automation

reason-machines/devtools-skills

Open the skill to read its SKILL.md instructions.

Productivity & Automation260 installs

autocli-web-scraping

reason-machines/devtools-skills

Open the skill to read its SKILL.md instructions.

Productivity & Automation257 installs

cli-anything-agent-native-software

reason-machines/devtools-skills

Open the skill to read its SKILL.md instructions.

257 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。

Productivity & Automation815.7K installs

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。

Productivity & Automation815K installs

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）。

Productivity & Automation815K installs

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.

Productivity & Automation814.8K installs

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。

Productivity & Automation814.1K installs