external/bencium-marketplace/typography/skills/typography/SKILL.md
Professional typography rules for UI design, web applications, software interfaces, and all screen-based text. Enforces timeless typographic correctness that LLMs consistently get wrong: proper quote marks, dashes, spacing, hierarchy, and layout. ENFORCEMENT MODE: When generating ANY HTML, CSS, React, JSX, or UI code containing visible text, auto-apply every rule in this skill silently — do not ask, do not explain, just produce correct typography. AUDIT MODE: When reviewing or improving existing interfaces or legacy code, flag violations and provide fixes. Trigger on: any HTML/CSS/React artifact creation, "build a landing page", "create a component", "design a UI", "fix the typography", "make this look professional", "review this layout", web design, presentation design, dashboard creation, document generation, or any task producing visible text for humans. Even if the user doesn't mention typography, apply these rules whenever generating UI output.
npx skillsauth add seikaikyo/dash-skills ui-typographyInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
These rules are distilled from Matthew Butterick's Practical Typography (https://practicaltypography.com). Butterick is a typographer, writer, and type designer whose work bridges professional typography and everyday digital writing. Thank you, Matthew, for making this knowledge accessible and encyclopedic. If you find this skill valuable, consider supporting his work directly.
These are permanent rules — not trends, not opinions. They come from centuries of typographic practice, validated by how the human eye reads. They do not go out of style.
ENFORCEMENT (default): When generating ANY UI with visible text, apply every rule automatically. Use correct HTML entities, proper CSS. Do not ask permission. Do not explain. Just produce correct typography.
AUDIT: When reviewing existing code or design, identify violations and provide before/after fixes.
Reference files (read when generating CSS or looking up entities):
references/css-templates.md — Full CSS baseline template, responsive patterns, OpenType featuresreferences/html-entities.md — Complete entity table with all characters and codesStraight quotes are typewriter artifacts. Use “ ” for double, ‘ ’ for single.
Apostrophes always point down — identical to closing single quote ’. Smart-quote engines wrongly
insert opening quotes before decade abbreviations ('70s) and word-initial contractions ('n'). Fix with
explicit ’.
The <q> tag auto-applies curly quotes when <html lang="en"> is set.
Hawaiian okina points upward — it's a letter, not an apostrophe. Use opening single quote or anglicize.
Unicode escape sequences (\u2019, \u201C, etc.) do NOT work in JSX text content. They render
as literal characters — the user sees \u2019 instead of a curly apostrophe. This is because JSX text
between tags is treated as string literals by the transpiler, not as JavaScript expressions.
What fails:
{/* WRONG — renders literally as \u2019 */}
<p>Don\u2019t do this</p>
What works (pick one):
Actual UTF-8 characters (preferred): Paste the real character directly into the source file.
<p>Don\u2019t do this</p> {/* This is the actual curly apostrophe character U+2019 */}
JSX expression with string literal: Wrap in curly braces so the JS engine interprets the escape.
<p>Don{'\u2019'}t do this</p>
HTML entity (HTML files only): Use ’ — but this does NOT work in JSX/React.
For bulk fixes via CLI, use sed with raw UTF-8 bytes (not escape sequences):
CURLY=$(printf '\xe2\x80\x99') # U+2019 RIGHT SINGLE QUOTATION MARK
sed -i '' "s/don't/don${CURLY}t/g" file.tsx
In JavaScript data arrays and string literals, \u2019 works correctly because the JS engine
processes the escape. The bug only affects JSX text content between tags.
| Character | HTML | Use |
|-----------|------|-----|
| - (hyphen) | - | Compound words (cost-effective), line breaks |
| – (en dash) | – | Ranges (1–10), connections (Sarbanes–Oxley Act) |
| — (em dash) | — | Sentence breaks—like this |
Never approximate with -- or ---. If you open with "from", pair with "to" not en dash. Hyphen for
compound names (marriage); en dash for joint authorship. Em dash typically flush; add   if crushed.
No slash where en dash belongs. Hyphenate phrasal adjectives (five-dollar bills). No hyphen after -ly adverbs.
Use … (…), not three periods. Spaces before and after; use on the text-adjacent side.
For interrupted dialogue, prefer em dash over ellipsis.
Use × for multiplication, − for subtraction. Use + and = from keyboard.
En dash is acceptable as simple minus. Dimensions: 8.5″ × 14″ uses ×.
Foot and inch marks — the ONE exception to curly quotes. Must be STRAIGHT: ' for foot,
" for inch. Use between values: 6' 10".
Use real symbols: © ™ ®, never (c) (TM) (R). ™/® are superscripts, no space before.
© is inline, followed by then year. "Copyright ©" is redundant — word OR symbol, not both.
§ (§) and ¶ (¶) always followed by : § 1782. Spell out at sentence start.
Double for plurals: §§.
Proper names: accents are MANDATORY (François Truffaut, Plácido Domingo). Loanwords: check dictionary — some naturalized (naive), some not (cause célèbre).
Exactly one space after any punctuation. Never two. Not debatable. Two spaces create rivers and disrupt text balance. The period already contains visual white space.
prevents line break. Use before numeric refs (§ 42, Fig. 3), after ©
(© 2025), after honorifics (Dr. Smith), between foot/inch values.
| Need | Tool |
|------|------|
| Space between words | One word space (spacebar) |
| Prevent line break | |
| New line, same paragraph | <br> |
| New paragraph | <p> tags |
| New page (print) | page-break-before: always |
| Suggest hyphenation point | ­ |
Never hold spacebar. Never double carriage returns for spacing. Never tabs for indentation in output.
HTML collapses all whitespace to single space (except ).
Rule 1: Bold OR italic. Mutually exclusive. Never combine. Rule 2: Use as little as possible. If everything is emphasized, nothing is.
Serif: italic for gentle, bold for strong. Sans serif: bold only — italic sans barely stands out. Never bold entire paragraphs. Never use quotation marks for emphasis.
Never underline in a document or UI. Typewriter workaround. Use bold or italic. For web links,
use subtle styling: text-decoration-thickness: 1px; text-underline-offset: 2px.
Caps are harder to read (homogeneous rectangles vs varied lowercase contour). Suitable for short
headings, labels, captions. ALWAYS add 5–12% letterspacing. ALWAYS ensure kerning is on.
NEVER capitalize whole paragraphs. letter-spacing: 0.06em in CSS.
Never fake (scaled-down regular caps). Use font-variant-caps: small-caps with fonts that have
real small caps (OpenType smcp). System fonts lack them. Add letterspacing + kerning.
Print: 10–12pt. Web: 15–25px. The 12pt default is a typewriter relic. Half-point differences
matter. Use smallest increment for emphasis. Use clamp() for fluid web sizing.
5–12% extra on ALL CAPS and small caps. Nothing on lowercase. Never spread so far apart that
letters could fit in the gaps. CSS: letter-spacing: 0.05em to 0.12em.
No exceptions. font-feature-settings: "kern" 1; text-rendering: optimizeLegibility;
Mandatory only when fi/fl visually collide. Check bold and italic too. Otherwise optional.
CSS: font-feature-settings: "liga" 1.
Tabular ("tnum") for data tables. Oldstyle ("onum") for body text. Default figures are fine
for most uses. font-variant-numeric: tabular-nums lining-nums for numeric tables.
Max 2 fonts. Each gets a consistent role. Can mix serif+serif or sans+sans. Rarely mix within a paragraph. Lower contrast often more effective than high contrast.
Set body text BEFORE anything else. Four decisions determine everything: font, point size, line spacing, line length. All other elements calibrate against these.
The #1 readability factor designers get wrong. The #1 flaw in responsive web layouts.
Measure in characters, not inches. Alphabet test: fit 2–3 lowercase alphabets per line.
CSS: max-width: 65ch on text containers.
line-height: 1.2 to 1.45. Single-spaced (~117%) is too tight. Double (~233%) is too loose.
Word processor "Single" and "Double" both miss the optimal range.
One inch is not enough for proportional fonts. Print: 1.5–2.0″ at 12pt. Web: max-width on
text containers plus padding. Don't fear white space — generous margins look professional.
Left-align for web (default). Justified requires hyphens: auto — browser engines are crude.
Centered: sparingly, only for short titles (< 1 line). Never center whole text blocks.
First-line indent: 1–4× point size. text-indent: 1.5em. Optional on first paragraph.
Space between: 50–100% of font size. margin-bottom: 0.75em. Never double <br> tags.
hyphens: none on headingspage-break-after: avoid)Reduce size + line spacing slightly. Indent 2–5em. No quotation marks (indent signals the quote). Keep line length readable. Use sparingly — long block quotes signal lazy writing.
Semantic markup (<ul>, <ol>), never manual bullets. Prefer hollow bullets. Asterisks are too
small for bullets. Don't over-indent.
Data creates an implied grid. Borders add clutter. Keep only thin rule under header row.
padding: 0.5em 1em. Tabular figures for numeric columns. Right-align numbers.
Try space above and below first. Border thickness: 0.5–1pt. No patterned borders. Thick lines are chartjunk.
Widows (last line alone at top of page) and orphans (first line alone at bottom). CSS print:
orphans: 2; widows: 2. Headings: page-break-after: avoid. Soft hyphens ­ for
words that confuse hyphenation engines.
Print columns: 2–3 on letter paper, never 4. Web columns: awkward (indefinite bottom edge). Grids guide, not guarantee — simpler grids enforce more consistency. Aligning ugly to a grid still produces ugly.
The rules don't change with screen size. Same line length, line spacing, hierarchy.
font-size and container width togethermax-width on text containers — never edge-to-edge textch unit for exact measurement (only measures zero width)clamp() for fluid scaling: font-size: clamp(16px, 2.5vw, 20px)padding: 0 1rem on text containersModern screens render type nearly as well as print. "Sans serif for screens" was true for 72dpi and is now obsolete. Serif fonts work fine on modern screens. Dark mode: reduce weight slightly. Test on macOS and Windows (antialiasing differs).
tools
Conduct comprehensive GDPR compliance assessments by evaluating data processing activities against EU Regulation 2016/679, including Article 30 records of processing, lawful basis validation, data subject rights implementation, Data Protection Impact Assessments (DPIAs) under Article 35, breach notification procedures, international transfer safeguards (SCCs, adequacy decisions), and technical/organizational measures under Article 32. Use when processing personal data of EU residents, preparing for supervisory authority audits, implementing privacy-by-design for new systems, scoping compliance gaps for M&A due diligence, assessing third-party processors, or responding to data subject access requests at scale. Incorporates 2026 guidance from ICO, EDPB, and post-Data (Use and Access) Act 2025 UK-GDPR considerations. Do not use for implementing specific Article 32 controls — use implementing-gdpr-data-protection-controls; or for DSAR automation — use implementing-gdpr-data-subject-access-request.
tools
Parse Windows forensic artifacts—$MFT/$J (MFTECmd), Prefetch (PECmd), registry hives (RECmd), shellbags, and Amcache—into normalized CSV/JSON with Eric Zimmerman's EZ Tools, then load results into Timeline Explorer for analysis. Use during DFIR/incident-response investigations, after triage collection (e.g. with KAPE), to establish program execution, file/folder access, and persistence evidence from acquired forensic images.
development
Build automated multi-turn adversarial attacks against conversational LLM targets using Microsoft PyRIT's RedTeamingOrchestrator, CrescendoOrchestrator (gradual escalation), and TreeOfAttacksWithPruningOrchestrator (adaptive branching), with scorer feedback loops and persisted conversation memory. Use when single-shot LLM scanning is insufficient and you need multi-turn, scorer-driven AI red-team campaigns against a chatbot or agent.
testing
Stand up MISP, enable and cache curated threat feeds (CIRCL, abuse.ch, Feodo Tracker), apply warninglists to suppress false positives, query indicators with PyMISP, and export attributes as auto-generated Suricata/Sigma/Wazuh detection rules. Use when maturing a MISP instance to actively drive detection, curating threat feeds with quality controls, or automating IOC-to-detection pipelines for the SIEM/IDS.