- Home
- Skills
- Frontend & Web Apps
- UI Recheck — Before & After
Works with the AI Agents you already use
UI Recheck — Before & After
Recheck the same UI states and widths. Export paired results with failed and unfinished checks in view.
$13
UI Recheck — Before & After
Example session with this skill installed
Use the included synthetic component snapshots and declared ordinary/long-content cases at 320 and 768 pixels. Run the matched before/after checks, retain the actual callback and focus-target results, and export a paired report. Preserve failures and missing observations. Do not edit product source, deploy anything or call the result an accessibility certification.
- Read your context and instructions
- Compiled the ui recheck
The retained Brave run on 31 October 2026 recorded all eight synthetic observations. At 320 pixels, the long-label toolbar's initial overflow fell from 255 pixels before to zero after. All eight observations passed their programmatic click-callback and focus-target checks. The failing baseline stayed visible, so the overall report correctly remained CHECKS_HAVE_FAILURES.
A real server stop/restart on the same port preserved the first observation; the resumed run added the seven missing cases without duplicates. Six negative observations recorded ERROR with diagnostics even though the expected callback matched. Separate boundary cases retained errors, unsupported states and failed assertions.
The complete HTML was materialized from the visible Prepared export and inspected in Brave. This does not establish a successful browser download. The example is synthetic; it does not prove native input, accessibility, production readiness or a customer's component. Exact-ZIP and extracted-package browser checks remain separate. See the package's worked example for the records and provenance.
Connects securely to your tools. The creator never sees your data.
About this skill
Compare the same UI states before and after a change. Get a paired report with each result and next check.
A fix at one width can leave a long label overflowing. Supply curated before and after snapshots of the real component. Declare the same states, widths and assertions for both. The workbench records text and state checks, initial horizontal overflow, programmatic click callbacks and focus-target identity. Failed and unfinished cases stay in view.
The included synthetic invoice-toolbar example uses ordinary and long labels at 320 and 768 pixels. It declares four matched pairs and eight observations. Export the recorded JSON and an offline HTML report with source identities and measured results.
After an interruption, reopen the exact local session in the same browser environment. Continue from its recorded observations. The saved restart command keeps the port needed for checkpoint recovery. Changed inputs or environment start a new run. Earlier evidence stays available.
Use the included Component Stress Lab to choose cases. Use the included UI Finish Pass for an authorized focused patch. Those free skills remain independently useful. The paid tool adds runnable comparison, matched records, checkpoint handling and export.
Materialize the package on a machine with Node.js 22+ and a modern browser. Supply a browser ESM adapter for the existing component. Product edits remain with your agent and project tools. Native input, focus visibility and screen-reader behavior need separate host checks. Host screenshots are attached separately; the export does not capture images or compare pixels.
Bring one component change and the states you want to repeat.
Launch offer: $13.00 USD one-time (at least 30% off) until 2026-11-17T08:47:17.967000Z (17 November 2026 at 09:47:17 CET, Europe/Amsterdam). The regular one-time price is $19.00 USD afterward.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Frontend & Web Apps.
Checks cover the declared local fixtures, states and widths. Text, state and overflow assertions run before the actions; click and focus checks are programmatic. Captured browser error and unhandled-rejection events force ERROR, even after a passing callback. Capture starts before adapter import and closes after a 100 ms post-action settle or the 4,500 ms child budget; errors after closure are excluded. Native pointer/keyboard behavior, focus visibility and screen-reader behavior need separate host evidence. This curated fixture tool is not a hostile-code sandbox or an accessibility/production-readiness certification. Exact-origin browser storage is required for checkpoint recovery. The runtime contract explains input bounds, isolation, recovery and report interpretation.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
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 16 days ago
- Software
Node.js 22+Modern browserHost file, shell and browser access
Permissions
Reads curated component snapshots and package files. Writes the chosen session receipt and report exports. Runs a local Node server and browser checks on 127.0.0.1 only; no external account or credential is required for the included example. Product edits need separate authorization.