Chrome Cdp

Chrome CDP

Lightweight Chrome DevTools Protocol CLI. Connects directly via WebSocket — no Puppeteer, works with 100+ tabs, instant connection.

Prerequisites

• Chrome with remote debugging enabled: open chrome://inspect/#remote-debugging and toggle the switch
• Node.js 22+ (uses built-in WebSocket)

Commands

All commands use ~/.claude/skills/chrome-cdp/~/.claude/skills/chrome-cdp/scripts/cdp.mjs. The <target> is a unique targetId prefix from list; copy the full prefix shown in the list output (for example 6BE827FA). The CLI rejects ambiguous prefixes.

List open pages

~/.claude/skills/chrome-cdp/scripts/cdp.mjs list

Take a screenshot

~/.claude/skills/chrome-cdp/scripts/cdp.mjs shot <target> [file]    # default: /tmp/screenshot.png

Captures the viewport only. Scroll first with eval if you need content below the fold. Output includes the page's DPR and coordinate conversion hint (see Coordinates below).

Accessibility tree snapshot

~/.claude/skills/chrome-cdp/scripts/cdp.mjs snap <target>

Evaluate JavaScript

~/.claude/skills/chrome-cdp/scripts/cdp.mjs eval <target> <expr>

Watch out: avoid index-based selection (querySelectorAll(...)[i]) across multiple eval calls when the DOM can change between them (e.g. after clicking Ignore, card indices shift). Collect all data in one eval or use stable selectors.

Other commands

~/.claude/skills/chrome-cdp/scripts/cdp.mjs html    <target> [selector]   # full page or element HTML
~/.claude/skills/chrome-cdp/scripts/cdp.mjs nav     <target> <url>         # navigate and wait for load
~/.claude/skills/chrome-cdp/scripts/cdp.mjs net     <target>               # resource timing entries
~/.claude/skills/chrome-cdp/scripts/cdp.mjs click   <target> <selector>    # click element by CSS selector
~/.claude/skills/chrome-cdp/scripts/cdp.mjs clickxy <target> <x> <y>       # click at CSS pixel coords
~/.claude/skills/chrome-cdp/scripts/cdp.mjs type    <target> <text>         # Input.insertText at current focus; works in cross-origin iframes unlike eval
~/.claude/skills/chrome-cdp/scripts/cdp.mjs loadall <target> <selector> [ms]  # click "load more" until gone (default 1500ms between clicks)
~/.claude/skills/chrome-cdp/scripts/cdp.mjs evalraw <target> <method> [json]  # raw CDP command passthrough
~/.claude/skills/chrome-cdp/scripts/cdp.mjs stop    [target]               # stop daemon(s)

Coordinates

shot saves an image at native resolution: image pixels = CSS pixels × DPR. CDP Input events (clickxy etc.) take CSS pixels.

CSS px = screenshot image px / DPR

shot prints the DPR for the current page. Typical Retina (DPR=2): divide screenshot coords by 2.

Tips

• Prefer snap --compact over html for page structure.
• Use type (not eval) to enter text in cross-origin iframes — click/clickxy to focus first, then type.
• Chrome shows an "Allow debugging" modal once per tab on first access. A background daemon keeps the session alive so subsequent commands need no further approval. Daemons auto-exit after 20 minutes of inactivity.