chrome-cdp

chrome-cdp is a skill for Claude Code, Codex from mattzcarey/shippie. It costs 91 tokens per session (1,700 once invoked), scanned A, original, MIT.

A browser-testing workflow that controls headless Chrome through the Chrome DevTools Protocol, a standard connection for inspecting and operating Chrome. It can navigate pages, interact with controls, inspect page structure, capture screenshots, and check results.

In plain words
What is it for?
Loading pages, clicking, entering form data, reading the page and accessibility tree, taking screenshots, and asserting that browser behavior or content is correct.
Why use it?
It lets you test a website through real browser actions without relying on a browser automation framework such as Playwright.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: positional $N argument.

Good fit Loading pages, clicking, entering form data, reading the page and accessibility tree, taking screenshots, and asserting that browser behavior or content is correct.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/mattzcarey/shippie/chrome-cdp
About the project

mattzcarey/shippie is an extensible code-review agent that examines code changes with developer tools and posts focused review comments. Teams use it to detect issues such as exposed secrets, inefficient code, bugs, and unhandled edge cases in pull-request workflows.

mattzcarey/shippie · 2,502 stars · on GitHub · shippie.dev

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add mattzcarey/shippie --skill chrome-cdp
Clone the repo
git clone --depth 1 https://github.com/mattzcarey/shippie

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for chrome-cdp

README.md
[![agentmods](https://agentmods.dev/badge/skills/mattzcarey/shippie/chrome-cdp/github.svg)](https://agentmods.dev/skills/mattzcarey/shippie/chrome-cdp)
Your own site
<a href="https://agentmods.dev/skills/mattzcarey/shippie/chrome-cdp"><img src="https://agentmods.dev/badge/skills/mattzcarey/shippie/chrome-cdp/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for chrome-cdp

Your own site · 80×15
<a href="https://agentmods.dev/skills/mattzcarey/shippie/chrome-cdp"><img src="https://agentmods.dev/badge/skills/mattzcarey/shippie/chrome-cdp.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 91 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,700 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector warn 7 Sept 2026
SkillSpector: 2 findings, up to medium

These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →

  • medium Rogue Agent · line 25
    Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
    Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
  • medium Rogue Agent · line 25
    Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
    Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00091 $0.01700
Opus 5 $0.00046 $0.00850
Sonnet 5 $0.00018 $0.00340
Haiku 4.5 $0.00009 $0.00170

Measured 13d ago against content hash ec196391dfb1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

chrome-cdp scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 13d ago.

The scan reads SKILL.md. This mod also ships 4 executable files (scripts/cdp-client.mjs, scripts/cdp.mjs, scripts/gen-test.mjs, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

src/skills/chrome-cdp/SKILL.md · 118 lines

How it starts

The opening of the file, as written. The whole thing — 118 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Driving Chrome over CDP (no MCP, no Playwright-for-driving)

scripts/cdp.mjs is a dependency-free CLI (Node 22+ built-in WebSocket). You launch headless Chrome from a bash step, then attach to it on a port. Every command is:

node .agents/skills/chrome-cdp/scripts/cdp.mjs --port $PORT <command> <target> [args]

where <target> is a targetId prefix from cdp list (copy the prefix shown, e.g. 18EB6379).

1. Launch ONCE per flow — survives across bash calls

Each bash tool call is a fresh process group, so a launched browser is only reachable later if it is detached. The helper handles that portably (setsid on Linux, nohup+disown on macOS), uses $CHROME_BIN, bakes in --no-sandbox --disable-dev-shm-usage (mandatory as root / in a container), and blocks until the CDP endpoint is ready. Use a unique port per flow (parallel-safe):

PORT=$(( 9222 + ${FLOW_INDEX:-0} ))
bash .agents/skills/chrome-cdp/scripts/launch-chrome.sh "$PORT"

2. Drive it. --port $PORT re-discovers the ws endpoint via /json/version each call —

no shell variable survives between bash calls, only the port does.

CDP="node .agents/skills/chrome-cdp/scripts/cdp.mjs --port $PORT"
T=$($CDP list | head -1 | awk '{print $1}')        # the open page's targetId prefix
$CDP nav  "$T" "$BASE_URL/login"
$CDP snap "$T"                                       # accessibility tree → derive getByRole locators
$CDP fill "$T" "input[name=email]" "[email protected]" # focus + select + insertText (React/Vue-safe)
$CDP fill "$T" "input[name=password]" "hunter2"
$CDP click "$T" "button[type=submit]"               # JS el.click() via eval — covers standard buttons/links
# $CDP clickxy "$T" 412 388                          # real Input.dispatchMouseEvent at CSS px, for
#                                                     #   real-input-only handlers (see `shot` DPR note)
$CDP eval "$T" "location.pathname"                  # assert post-conditions
$CDP shot "$T" /tmp/f$PORT-01.png                   # then `read` the PNG (vision) to assert visually
$CDP evalraw "$T" "DOM.getDocument" '{}'             # raw-CDP escape hatch (method + JSON params)

Read the full file on GitHub · 118 lines

Files

What ships with it

4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 13d ago First seen · 118 lines · 91 tokens per session scan A ec196391dfb1

Subscribe to this mod's changes

chrome-cdp is a skill published in the GitHub repository mattzcarey/shippie (2,502 stars, last pushed yesterday), licensed MIT. It adds 91 tokens to every session and 1,700 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

ring:checking-frontend-quality

Checking frontend quality against changed UI via ring:qa-frontend in accessibility, visual, e2e, or performance mode and aggregating pass/fail verdicts. Use when a frontend change needs standalone a11y, visual-snapshot, Playwright e2e, or Lighthouse/Core-Web-Vitals validation outside the dev cycle. Skip for…

LerianStudio/ring · 102 tokens

design

Intent-first UI design with conscious, traceable design decisions. Persists decisions in .interface-design/ for cross-session consistency. Includes domain exploration, design system generation (system.md + system.json), per-component construction with mandatory checkpoints, and craft validation tests. Modes: init…

greglas75/zuvo · 90 tokens

write-e2e

Generate Playwright E2E tests from codebase analysis. Discovers routes, scores user flows by criticality, writes .spec.ts files that assert causality, and reports what was actually proven: GENERATED, STATICCHECKED, VERIFIEDLOCAL or VALIDATEDLIVE. Modes: --scope , --flow , --output , --base-url , --max-flows N, --live…

greglas75/zuvo · 101 tokens

e2e-testing

Playwright E2E testing patterns — Page Object Model, configuration, CI/CD integration, artifact management, flaky test strategies, PLUS the decision-time patterns that come before the first test is written (black-box helper scripts, static-vs-dynamic routing, multi-server orchestration, reconnaissance-then-action…

Nmor/the-claude-council · 70 tokens

browser-testing-with-devtools

Tests in real browsers. Use when building or debugging anything that runs in a browser. Use when you need to inspect the DOM, capture console errors, analyze network requests, profile performance, or verify visual output with real runtime data via Chrome DevTools MCP.

manu14357/zskills · 57 tokens

test-conversion-reviewer

Reviewer for converting unit tests to browser tests for Project Bedrock. Invoke this when the user needs to review their work while removing complex Browser dependencies from tests.

chromium/chromium · 37 tokens