Cross Browser Typography Qa logo

Cross Browser Typography Qa

Community
AtlasOmnia
cross-browser-typography-qa

cross-browser-typography-qa — Diagnose and verify web typography rendering defects across Chromium, WebKit, and native Safari, including clipped glyphs, broken descenders, wrapping, font metrics, gradient text, and live-cache mismatches.

Overview

PublisherAtlasOmnia
Repositorydonna-starter
Skill namecross-browser-typography-qa
Stars
131
Forks
9
Bundled files
4
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 4 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by AtlasOmnia on GitHub. Read the source before you install it.

Installation

Install the Cross Browser Typography Qa AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/AtlasOmnia/donna-starter/tree/main/skills/software-development/cross-browser-typography-qa
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/AtlasOmnia/donna-starter.git /tmp/donna-starter
mkdir -p .claude/skills
cp -r /tmp/donna-starter/skills/software-development/cross-browser-typography-qa .claude/skills/cross-browser-typography-qa
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Cross Browser Typography Qa in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Cross Browser Typography Qa on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Cross Browser Typography Qa is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Cross-Browser Typography QA

Purpose

Use this skill when web text looks clipped, flattened, wrapped incorrectly, misaligned, or materially different between Chromium and WebKit/Safari. Treat text rendering as both a geometry problem and a visual paint problem: passing DOM bounds does not prove that every glyph is intact.

Triggers

  • Lowercase descenders such as g, j, p, q, or y appear cut off.
  • Gradient or transparent text looks flattened at the baseline.
  • A heading fits within its element but still looks clipped.
  • Safari and Chromium disagree on line height, wrapping, button labels, or heavy font weights.
  • A deployed fix appears absent despite a successful upload or publish message.

Workflow

1. Reproduce the exact defect

  1. Record the browser family, viewport, font family, weight, size, and exact text.
  2. Use a string containing ascenders, rounded forms, punctuation, and descenders.
  3. Capture a focused screenshot at the reported viewport before changing CSS.
  4. Inspect the actual glyphs, not only the element rectangle.

2. Inspect layout and paint

Read computed values for:

  • font-family, font-size, font-weight, and line-height;
  • display, overflow, padding, transforms, and fixed heights;
  • background-clip, -webkit-background-clip, and -webkit-text-fill-color;
  • loaded font status and fallback-font substitution;
  • the stylesheet URL/cache marker actually active in the page.

Use Range.getClientRects() for rendered-line counts and getBoundingClientRect() for containment, but never treat either as proof that glyph paint is complete.

3. Repair the cause

Choose the smallest stable repair:

  • Treat a user's “is it meant to be sized like that?” reaction as a real visual-acceptance signal, not merely a request to confirm CSS intent. First verify the actual browser window bounds and distinguish normal viewport rendering from browser zoom or an accidentally narrow layout. If the viewport is normal but the display type dominates the first screen, make a bounded desktop-only scale/spacing adjustment while preserving explicit mobile overrides.
  • For a hero-scale correction, adjust the heading clamp() and hero padding-block together; shrinking only the type can leave the composition oddly empty. Add an exact contract assertion before the CSS change so the intended scale is durable, prove RED, then rerun five-width Chromium and native-Safari geometry plus visual inspection.
  • When the user asks to make a small, even-numbered product set “even,” prefer an equal two-column grid over editorial 7/5 nth-child spans. Encode the parity in a RED contract, remove obsolete asymmetric and breakpoint-reset rules, retain the one-column mobile override, update the stylesheet cache marker, and visually inspect paired card alignment—not only grid geometry.
  • Increase the line box on the text-bearing element when font paint exceeds a tight line box.
  • Remove fixed heights or hidden overflow that cut off glyphs.
  • Avoid synthetic or unsupported font weights when they distort metrics.
  • Rework pathological wrapping with width, balance, or explicit block structure.
  • When a semantic hyphenated term breaks awkwardly, protect only that token with an inline span using white-space: nowrap; do not force the entire heading onto one line. Keep the ordinary hyphen when exact-copy tests, translation keys, or search indexing depend on the literal string.
  • For background-clip: text, prefer line-height correction over outside padding; outside padding can enlarge the box without enlarging the painted glyph mask.
  • If a decorative gradient remains unstable, use a solid accessible text color rather than preserving a fragile effect.
  • When prose interleaves raw text with an inline emphasis element inside display: flex, wrap the entire sentence in one inline-flow child (for example, <span>Press <strong>Start Session</strong> ...</span>). Otherwise the browser can promote text fragments and the emphasis element to separate anonymous flex items, making intended spaces appear missing or incorrectly distributed. Verify the actual rendered glyph spacing in the packaged app, not only the JSX or DOM text.

Mixed-inline-content spacing regression

Treat a screenshot showing PressStart Sessionto...-style spacing as a layout-structure defect before changing font metrics. Inspect the parent display mode and child structure; if text nodes surround an inline child in a flex container, use one wrapper element rather than inserting arbitrary literal spaces or changing word-spacing. Rebuild the production renderer and capture the exact installed process/window after replacement so stale or pre-fix surfaces cannot be mistaken for current evidence.

4. Verify locally

Test at minimum:

  • 375 and 390 px mobile;
  • the exact reported width;
  • 1024 and 1440 px;
  • Chromium and WebKit/Safari.

Before accepting any narrow screenshot, prove that the CSS viewport is actually narrow. Some headless Chromium invocations keep a minimum layout viewport around 500 CSS px and crop a --window-size=390 screenshot to 390 physical pixels. In-page, read innerWidth, document.documentElement.clientWidth, and document.documentElement.scrollWidth. If innerWidth does not equal the requested width, use CDP Emulation.setDeviceMetricsOverride, assert the width again, and only then inspect the screenshot. When overflow is real, inspect element bounding boxes and repair intrinsic grid/flex sizing with minmax(0, 1fr) or min-width: 0; do not hide the defect with overflow-x: hidden before identifying its cause.

Require:

  • naturally shaped descenders;
  • no unexpected wrapping or overlap;
  • balanced spacing below the text;
  • no horizontal overflow or console errors;
  • one intended semantic heading where applicable.

When visual evidence follows an interaction, make UI state explicit. For a mobile menu, capture the open state after aria-expanded="true" and visible-link checks, then close it and wait for both aria-expanded="false" and the menu to become non-displayed before taking the clean layout screenshot. A successful click or DOM-count pass does not prove the screenshot captured the intended state.

5. Verify production

  1. Add a new stylesheet cache marker when deploying changed CSS.
  2. Load a fresh production URL, preferably with a harmless query marker.
  3. Read back the live stylesheet URL and computed values.
  4. Capture a fresh live screenshot and inspect the originally failing glyphs.
  5. Treat deployment success as transport evidence only; close the defect from live rendering evidence.

Common pitfalls

  • Geometry-only false pass: The element is on-screen and overflow: visible, but the glyph mask is clipped.
  • Padding-only false fix: padding-bottom increases element height while gradient-painted descenders remain flattened.
  • Stale-tab confusion: An old tab continues using a previous cache-busted stylesheet after a new deployment.
  • Local-only acceptance: A local screenshot passes, but production serves an older stylesheet.
  • Browser-family substitution: Chromium at Safari-like dimensions is not a Safari/WebKit acceptance gate.
  • Screenshot scale blindness: Full-page screenshots can hide a one- or two-pixel baseline defect; inspect a focused crop or zoomed view.

Verification snippet

js
const el = document.querySelector('.gradient-text');
const cs = getComputedStyle(el);
const rect = el.getBoundingClientRect();
({
 stylesheet: document.querySelector('link[rel="stylesheet"]')?.href,
 fontFamily: cs.fontFamily,
 fontWeight: cs.fontWeight,
 lineHeight: cs.lineHeight,
 overflow: cs.overflow,
 backgroundClip: cs.backgroundClip,
 webkitTextFillColor: cs.webkitTextFillColor,
 rect: { top: rect.top, bottom: rect.bottom, height: rect.height }
});

This snippet confirms state; it does not replace visual inspection.

  • references/gradient-text-descenders.md — focused diagnosis and repair pattern for clipped descenders in gradient display text.
  • references/responsive-heading-wrap-and-ui-state.md — contract-first repair for awkward semantic-hyphen wrapping plus native-Safari open/closed screenshot-state acceptance.
  • references/desktop-hero-scale-calibration.md — bounded desktop hero resizing from a real-browser acceptance signal, with RED contract, cache marker, five-width QA, and visible-tab refresh.
  • references/responsive-product-card-grid-parity.md — user-driven conversion from asymmetric editorial spans to equal card rows, with RED contract, cache-busting, cross-browser QA, and local-Git closeout.

Public support files

  • references/desktop-hero-scale-calibration.md
  • references/gradient-text-descenders.md
  • references/responsive-heading-wrap-and-ui-state.md
  • references/responsive-product-card-grid-parity.md

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Cross Browser Typography Qa AI skill do?

cross-browser-typography-qa — Diagnose and verify web typography rendering defects across Chromium, WebKit, and native Safari, including clipped glyphs, broken descenders, wrapping, font metrics, gradient text, and live-cache mismatches.

Why use Cross Browser Typography Qa on TypingMind?

Because you install it once and use it with any model. Cross Browser Typography Qa is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Cross Browser Typography Qa in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/AtlasOmnia/donna-starter/tree/main/skills/software-development/cross-browser-typography-qa. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Cross Browser Typography Qa?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Cross Browser Typography Qa?

As many as you like. As long as a model supports skills, you can use Cross Browser Typography Qa with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Cross Browser Typography Qa AI skill free?

Yes. It is published on GitHub by AtlasOmnia under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇