mgifford/accessibility-skills/plain-language
Claude Code and agent skill by mgifford/accessibility-skills
- community
Checking that the SKILL.md file is still in the repository.
Load this skill whenever the project involves writing, editing, or reviewing content for a general audience, including UI copy, help text, error messages, documentation, and instructions. Under no circumstances use jargon, complex sentence structures, or unexplained acronyms without providing plain alternatives. Absolutely always target a reading level accessible to the broadest possible audience. Load alongside content-design/SKILL.md for all content work.
Install
npx vskill@latest install mgifford/accessibility-skills/plain-languageSKILL.md
Plain Language Accessibility Skill
Canonical source:examples/PLAIN_LANGUAGE_ACCESSIBILITY_BEST_PRACTICES.mdinmgifford/ACCESSIBILITY.mdThis skill is derived from that file. When in doubt, the example is authoritative.
Apply these rules when writing, editing, or reviewing web content. Only load this skill if the project involves content creation or editing.
This skill complements content-design/SKILL.md — load both for content work. Plain language is the deeper implementation; content design is the structural pattern.
Core Mandate
Plain language means writing so that people can find what they need, understand it, and act on it the first time they read it. It is not about dumbing content down — it is about respecting the reader's time and removing unnecessary barriers.
Plain language is the primary accessibility strategy for:
- Users with cognitive disabilities, learning disabilities, and intellectual disabilities
- Users with low literacy or whose first language is not the page language
- Users who are stressed, fatigued, or in crisis
- Users accessing content via screen reader, where re-reading is slow
- Users in reader mode or using text-to-speech, where complex syntax breaks down
WCAG 3.1.5 Reading Level (AAA) targets Grade 8 or lower for general audiences. Plain language is also an Opquast requirement (rules 5, 7) and aligns with the Web Sustainability Guidelines — concise content reduces data transfer and reading time.
Severity Scale (this skill)
| Level | Meaning | |
|---|---|---|
| Critical | Content is entirely incomprehensible to the target audience; essential instructions cannot be followed | |
| Serious | Dense jargon or complex sentence structure creates a significant barrier for users with cognitive disabilities | |
| Moderate | Reading level above Grade 8; abbreviations unexpanded; passive voice dominates | |
| Minor | Opportunities to shorten or clarify that do not create a barrier |
Assistive Technology Context
Plain language matters differently across AT:
| AT | How plain language helps | |
|---|---|---|
| Screen readers (NVDA, JAWS, VoiceOver) | Short sentences reduce cognitive load when re-listening; heading hierarchy aids navigation | |
| Reader Mode (Firefox/Edge/Safari) | Strips visual hierarchy — plain language structure must carry meaning alone | |
| Edge Read Aloud / TalkBack TTS | Long sentences with many clauses are harder to parse when heard | |
| Voice Control | Clear, short headings and labels are easier to speak as navigation targets | |
| Screen magnification | Users see small portions at a time — front-loaded key information reduces scrolling | |
| Cognitive accessibility tools (e.g., Helperbird, Recite Me) | These reformat content; simple structures reformat better than complex ones |
Critical: Structure Before Prose
Plain language begins with structure. A well-structured page with mediocre prose is more accessible than beautiful prose with poor structure.
One topic per page; one idea per paragraph
Front-load the key information
Put the conclusion first, then the evidence. Readers should not have to reach the end of a paragraph to understand what it is about.
Inverted pyramid: most important → supporting detail → background
This structure serves screen reader users (who navigate by heading), search engine users, and users in crisis who need the answer immediately.
Serious: Sentence and Paragraph Length
Long sentences and paragraphs are a Serious barrier for users with cognitive disabilities, dyslexia, and low literacy.
Sentences:
- Target 15–20 words maximum for most sentences
- One main clause per sentence; move subordinate clauses to their own sentence
- Remove hedges and qualifications where possible: "It should be noted that" → remove it
Paragraphs:
- 3–5 sentences maximum for general web content
- One main idea per paragraph
- Use blank lines between paragraphs (not just indentation)
Serious: Active Voice
Passive voice hides the actor and increases cognitive load. Use active voice unless there is a deliberate reason not to.
When to use passive voice deliberately:
- The actor is truly unknown: "Three people were injured."
- The actor is irrelevant: "The form was designed in 2019."
- You want to soften a negative: "Mistakes were made." (rare, and be cautious)
Serious: Plain Words
Use the word your reader already knows. Technical terminology is appropriate for expert audiences; general audiences need plain equivalents.
| Instead of | Use | |
|---|---|---|
| utilise | use | |
| commence | start | |
| endeavour | try | |
| prior to | before | |
| in the event that | if | |
| at this point in time | now | |
| facilitate | help | |
| demonstrate | show | |
| subsequently | then / after | |
| approximately | about |
Domain-specific terms (legal, medical, technical): define them on first use inline, not in a separate glossary unless the audience is expert.
Moderate: Reading Level
Target Grade 8 or lower (Flesch-Kincaid) for general public content. Above Grade 10 is Moderate; above Grade 12 is Serious for public-facing content.
Tools for checking reading level:
- Hemingway Editor — highlights complex sentences, passive voice, grade level
retext-readability— CI-integrable readability linter- Microsoft Word's built-in Flesch-Kincaid score (Review → Check Accessibility)
Grade level is a guide, not an absolute rule. Legal or regulatory text may require precision that raises the grade level — acknowledge this and provide a plain language summary alongside the formal text.
Moderate: Abbreviations and Acronyms
Unexpanded abbreviations are a Moderate barrier — screen readers may mispronounce them; new readers do not know them.
Rule: Expand on first use on every page (not just the first page of the site).
<abbr title="..."> provides the expansion to AT and on hover for sighted users. Screen readers handle <abbr> inconsistently — the expansion in the title attribute is read by some but not all AT. Always write the expansion in text on first use regardless of whether you use <abbr>.
Moderate: Lists vs Prose
Use lists for three or more parallel items. Embedding lists in prose hides the structure and increases cognitive load.
Use <ul> for unordered lists, <ol> for steps or sequences that must be followed in order.
Moderate: Instructions and Sequences
Number steps. Place the action before the context (not after).
Minor: Tone and Inclusive Language
- Address the reader as "you"; address your organisation as "we"
- Gender-neutral language throughout: "they" as singular is grammatically
accepted and widely supported; avoid "he or she"
- Avoid ableist language: do not use disability as metaphor
("turning a blind eye", "deaf to their concerns", "stands alone")
- Avoid culture-specific idioms that may not translate: "ballpark figure",
"boil the ocean", "hit the ground running"
Minor: Page Titles and Headings
- Page
<title>must identify the page and the site: "About Us | Acme Corp" - One
<h1>per page; must match or closely reflect the<title> - Headings must be descriptive — "Introduction" alone is not descriptive;
"Introduction to plain language principles" is
- Do not use headings for visual styling — use CSS classes instead
Testing Checklist
Run these checks before publishing:
- Readability score: Flesch-Kincaid Grade 8 or lower for general content
- Sentence scan: no sentence over 25 words without good reason
- Passive voice scan: Hemingway Editor highlights these in green
- Abbreviation check: every abbreviation expanded on first use
- Plain word check: replace any word from the "instead of" table above
- Reader Mode test: open in Firefox or Safari Reader Mode — does it make
sense without CSS? Are headings correct? Is content in logical order?
- Read aloud test: use Edge Read Aloud or macOS VoiceOver — confusing
sentences sound obviously wrong when spoken
Definition of Done Checklist
- [ ] Reading level: Grade 8 or lower (Flesch-Kincaid) for general audiences
- [ ] Sentences: ≤20 words average; no sentence over 30 without good reason
- [ ] Active voice used throughout; passive voice only when deliberate
- [ ] Key information front-loaded in each paragraph
- [ ] All abbreviations and acronyms expanded on first use on every page
- [ ] Technical terms defined inline on first use
- [ ] Lists used for three or more parallel items
- [ ] Numbered steps for sequential instructions
- [ ] Inclusive, gender-neutral language
- [ ] Page
<title>in "Page | Site" format - [ ] One
<h1>; headings descriptive and hierarchical - [ ] Reader Mode test passed
- [ ] Read Aloud test passed
Key WCAG Criteria
- 3.1.1 Language of Page (A) —
langattribute on<html> - 3.1.2 Language of Parts (AA) —
langon elements in a different language - 3.1.5 Reading Level (AAA) — Grade 8 or lower target for general audiences
- 1.3.1 Info and Relationships (A) — heading hierarchy, list markup
- 2.4.2 Page Titled (A) — descriptive page title
References
This is the start of the file. Read the full SKILL.md on GitHub.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/plain-language/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/mgifford/accessibility-skills/plain-language)