Chrome DevTools MCP is an MCP server that lets coding agents control and inspect a live Chrome browser for automation, debugging, screenshots, network analysis, console inspection, and performance tracing. It is for developers who want AI coding assistants to test and diagnose web applications, and the catalogue add-ons provide workflows for using those browser tools.
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.
npx skills add ChromeDevTools/chrome-devtools-mcp --skill cookie-debugginggit clone --depth 1 https://github.com/ChromeDevTools/chrome-devtools-mcpWrote 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.
[](https://agentmods.dev/skills/chromedevtools/chrome-devtools-mcp/cookie-debugging)<a href="https://agentmods.dev/skills/chromedevtools/chrome-devtools-mcp/cookie-debugging"><img src="https://agentmods.dev/badge/skills/chromedevtools/chrome-devtools-mcp/cookie-debugging/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.
<a href="https://agentmods.dev/skills/chromedevtools/chrome-devtools-mcp/cookie-debugging"><img src="https://agentmods.dev/badge/skills/chromedevtools/chrome-devtools-mcp/cookie-debugging.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Snyk warn
- NVIDIA SkillSpector warn
- Socket pass
SkillSpector: 1 finding, 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 Prompt Injection · line 20 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00071 | $0.02286 |
| Opus 5.5 | $0.00028 | $0.00914 |
| Sonnet 5.5 | $0.00014 | $0.00457 |
| Haiku 4.5 | $0.00007 | $0.00229 |
Grade A, and why
cookie-debugging scanned grade A with 1 finding 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 4d ago.
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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- Call `evaluate_script` with `() => fetch(window.location.href)` Copies of this mod
2 near-identical copies found in the catalogue:
- cookie-debugging — 100% identical, 13 lines differ
- cookie-debugging — 97% identical, 2 lines differ
How it starts
The opening of the file, as written. The whole thing — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Core Concepts
Page Targeting: Page-scoped tools (list_network_requests, get_network_request, list_console_messages, evaluate_script, navigate_page, take_snapshot, click, lighthouse_audit, close_page, etc.) require a pageId parameter. Retrieve available page IDs using list_pages or from new_page.
HttpOnly vs Client-Side Storage
Cookies marked HttpOnly cannot be accessed or modified by client-side JavaScript (cookieStore or document.cookie). However, the browser automatically attaches active HttpOnly cookies to outgoing HTTP request headers (Cookie).
- To inspect current
HttpOnlyvalues: Look at theCookierequest header of any outgoing HTTP request viaget_network_request. - To inspect how cookies were created or configured: Look at the
Set-Cookieresponse header of login/auth responses. - To inspect non-
HttpOnlycookies: Useevaluate_scriptwith the moderncookieStoreAPI (async () => await cookieStore.getAll()).
Session Strategy: Live Tab vs Isolated Context
Choose the right session environment to avoid state contamination (e.g., residual analytics or auth tokens):
| Strategy | When to Use | Setup / Teardown |
|---|---|---|
| Live Tab (Active Page) | Diagnosing an active user session, live 401/403 error, or current state. | Operates on the user's existing page; get its pageId from list_pages. |
Clean-Slate (isolatedContext) |
Testing cookie consent banners, first-time visits, or zero-cookie guarantees. | Call new_page with a unique isolatedContext (e.g. "consent-audit-1"). When finished, call close_page. |
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.
- 4d ago Changed f0f0daa1d75d
- 12d ago Changed · +3 lines 136f22496def
- 1mo ago First seen · 150 lines · 71 tokens per session scan A eeb526491ce5
cookie-debugging is a skill published in the GitHub repository ChromeDevTools/chrome-devtools-mcp (53,865 stars, last pushed yesterday), licensed Apache-2.0. It adds 71 tokens to every session and 2,286 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-02.
Other skills, from other repositories
devcdp
How to debug a running web app with the DevCDP tools - connect to a tab, reload to capture, drive the page with the ui tools, set breakpoints that bind, and hand off to the user. Use whenever DevCDP tools are available and the task involves a web page, the browser, console or network output, or a front-end bug.
brave-devtools
Uses Brave DevTools via MCP for efficient debugging, troubleshooting and browser automation. Use when debugging web pages, automating browser interactions, analyzing performance, or inspecting network requests. This skill does not apply to --slim mode (MCP configuration).
cookie-debugging
Uses Brave DevTools MCP for inspecting, debugging, and testing cookies, session state, authentication issues, and cookie consent compliance. Use when diagnosing 401/403 errors, authentication redirects, session expiration, Cookie/Set-Cookie header issues, cookie banner consent conformance, or third-party…
a11y-debugging
Uses Brave DevTools MCP for accessibility (a11y) debugging and auditing based on web.dev guidelines. Use when testing semantic HTML, ARIA labels, focus states, keyboard navigation, tap targets, and color contrast.
surf
Control Chrome browser via CLI for testing, automation, and debugging. Use when the user needs browser automation, screenshots, form filling, page inspection, network/CPU emulation, DevTools streaming, or AI queries via ChatGPT/Gemini/Perplexity/Grok/AI Studio.
chrome-bridge
Recipes for the user's real, logged-in Chrome through the chrome-bridge MCP tools and the zero-token chrome-bridge CLI. Use when a browser task needs a multi-step recipe — a form "until the email arrives", a checkout with a test card, a pixel/GTM check, a cookie-banner audit, redirects after a migration, a page…