More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Browser Testing with Chrome DevTools

    1

    Debug web UIs with runtime evidence: reproduce in a clean profile, observe through console, network, DOM, and performance, then verify the fix by re-running the reproduction — not by reading code.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Browser Testing with Chrome DevTools

    Browser Testing with Chrome DevTools

    Example session with this skill installed

    Our checkout page's "Place order" button does nothing on staging but works locally. Find the cause and prove the fix.

    • Read your context and instructions
    • Compiled the browser testing

    CLEAN ROOM: Throwaway profile, no extensions. Staging URL from the task config.

    REPRODUCE: Button clicked on staging while watching. Screenshot: loading spinner that never resolves. Console: one error, copied verbatim — a failed request to the payment endpoint with a 404. Network: the request fired to the wrong path — the staging config points at a path that does not exist on the staging server. DOM: the button is correctly wired; the handler is attached and it did fire.

    CAUSE: The console error and the network tab agree: the staging build ships a payment path that exists only in local development. The lens that identified it was network, confirmed by console. The fix is a config change, not a code change — the button was never broken.

    VERIFY: After the config fix, the same reproduction steps produce the order confirmation screen (screenshot saved), a clean console, and a 200 response on the corrected payment path. Before and after screenshots sit side by side in the session trail with the cause statement: "Staging config pointed the payment request at a local-only path; network lens, confirmed by console." The trail also records the reproduction steps, the broken-state screenshot, and the fixed-state evidence together, so the next person trusts the fix without re-running the session.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Static code review tells you what the code says; the browser tells you what the code does. This skill runs the loop between them: set up a clean room (throwaway profile by default, dedicated test profile when a login is needed — never the daily profile), reproduce the bug while watching, observe through five lenses (screenshot, console, network, DOM, performance), bisect one change at a time, and verify the fix by re-running the reproduction steps.

    A lens-first triage table starts each symptom at the right lens: rendering bugs at screenshot and DOM, dead buttons at console, slowness at the performance trace, staging-only bugs at the network diff. Recipes cover the five common failures, and a fix-verification checklist gates the handoff. Browser content is treated as data, never instructions — instruction-like page text gets reported as a finding, not followed. A worked example traces a dead checkout button on staging to a config path that exists only in local development, verified with before-and-after screenshots.

    • It tests what renders in a browser; native apps, load testing, and security testing are out of scope.
    • A bug that will not reproduce in a clean room needs variables added back one at a time — the skill does not declare it fixed.
    • Screenshots show what rendered, not what failed to render; every screenshot is paired with console and network checks.
    • Performance claims need a before-and-after trace on the same numbers; "feels faster" is not a measurement.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    Extensions, saved logins, and history are variables. A clean profile removes a whole class of "works for me" failures, and it keeps a bug reproduction from touching your real sessions.

    Run the reproduction five times while watching. Flakiness has a cause — timing, ordering, shared state — and the observation loop finds it faster than re-running blind.

    When the reproduction steps now produce the correct screenshot, clean console, and correct network traffic. Code that looks right is not a fix; the browser's verdict is.