Internal prototype QA helper. Use only after a Product Design prototype, URL-to-code build, or image-to-code build has a source visual target and a rendered implementation to compare before handoff. Do not use for broad UX critique, design critique, product audits, or flow reviews; route those user-facing requests to audit.
78%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
Passed
No findings from the security scan
Fix and improve this skill with Tessl
tessl review fix ./plugins/product-design/skills/design-qa/SKILL.mdUse this internal helper to compare a prototype's source design against the rendered implementation before handoff.
Do not use this skill for broad UX critique, design critique, product audits, or flow reviews. Use audit for those user-facing requests.
Use this skill before every Product Design build handoff.
A passing QA run requires both:
If either artifact cannot be opened, captured, or compared, write design-qa.md with final result: blocked and name the blocker. Do not let the build skill hand off as done.
Follow critical-overrides.
Compare the intended design to the implementation as a product-quality reviewer, not as a generic aesthetic critic. The output must be a prioritized fix list grounded in evidence from both artifacts.
Do not write the QA review from memory, code, or file paths alone. Open or capture both the source design and the implementation first, then compare what is actually visible.
Do not pretend separate image views are side-by-side comparison. Put the source image and the implementation screenshot together in the same comparison input, then judge the visible differences from that combined input.
Design QA is an iteration loop. The first comparison may pass only when it finds no actionable P0/P1/P2 differences and no visual fixes are made in response.
When a comparison finds any P0/P1/P2 issue:
A later pass must identify the earlier findings, the fixes made, and the post-fix visual evidence. Build, dependency, lint, deployment, and preview troubleshooting do not count as design-QA iterations.
Identify the comparison target.
Capture evidence.
mobile-app template implementations, capture the app viewport itself, not the whole browser page, desktop canvas, or surrounding phone stage. Use data-testid="device-screen" / [data-phone-screen] for content-only comparisons; use data-testid="phone-frame" only when the source visual includes the device bezel.mobile-app template, run npm run check:runtime. Treat a failure as blocking. Do not bypass the check or accept a rasterized or duplicated status bar, device bezel, or home indicator as app content.Normalize before comparing.
mobile-app template, force or verify a 1:1 phone-screen capture before judging fidelity. The template scales the device down to fit small browser viewports; use a large enough Playwright viewport for scale 1 and verify [data-phone-screen] measures 393 x 852 CSS px before capture. If the screen measures smaller, the screenshot is scaled and is not valid for 1:1 comparison.const page = await browser.newPage({
viewport: { width: 1400, height: 1200 },
deviceScaleFactor: 1,
});
await page.goto("http://127.0.0.1:8796");
const screen = page.getByTestId("device-screen");
await screen.waitFor({ state: "visible" });
const box = await screen.boundingBox();
if (!box || Math.abs(box.width - 393) > 1 || Math.abs(box.height - 852) > 1) {
throw new Error(`Expected unscaled mobile screen at 393 x 852, got ${box?.width} x ${box?.height}`);
}
await screen.screenshot({ path: "implementation-mobile-screen.png" });@2x or otherwise double-density, compare against a same-size normalized copy: either downsample 786 x 1704 source captures to the 393 x 852 CSS target, or capture the implementation at the same density and compare equal pixel dimensions. Record the source pixels, implementation pixels, CSS viewport, and deviceScaleFactor in design-qa.md.Compare at the right level of detail.
design-qa.md.Review systematically.
Produce a fix-oriented QA report.
Every QA report must explicitly evaluate these surfaces:
P0: Blocks core use, severe accessibility failure, broken layout, or impossible task.P1: Major design mismatch or usability regression likely to be noticed by users.P2: Moderate visual drift, inconsistent state, responsive issue, or fixable polish gap.P3: Minor refinement that improves fidelity but does not block acceptance.Treat viewport overflow that hides persistent app controls, and mismatches that materially change above-the-fold content, major-region proportions, text wrapping, or interface density, as P2 or higher.
Use this structure unless the user asks otherwise:
**Findings**
- [P1] Short issue title
Location: screen/component/selector/file if known.
Evidence: design does X, implementation does Y.
Impact: why this matters.
Fix: concrete change.
**Open Questions**
- Any ambiguity about intentional deviations, unavailable states, or missing artifacts.
**Implementation Checklist**
- Ordered fixes that can be executed directly.
**Follow-up Polish**
- P3 refinements that can improve fidelity after handoff.If there are no substantive mismatches, say that clearly and list any residual test gaps.
When this skill is used before handoff, save the latest QA report as project-root design-qa.md.
design-qa.md must include:
For ChatGPT Work Mode builds, design-qa.md must include the browser-rendered implementation screenshot, viewport, primary interactions tested, console errors checked, and final result. If browser-rendered evidence is missing, final result is blocked.
final result must be exactly passed or blocked.
Use passed when there are no actionable P0/P1/P2 findings. P3 findings may remain as follow-up polish.
Use blocked when actionable P0/P1/P2 findings remain and name the blocker.
Return the file path with the QA report.
openai/plugins
d420d13
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.