Agent skill

Web Platform Baseline

by WrongStack in WrongStack/WrongStack

Use this skill before asserting that a CSS, HTML or accessibility capability is available, unavailable, or the right tool — it carries dated, refreshable platform facts and refuses to let stale…

MITAuto-check passedFrontend & Design

Install Web Platform Baseline

skills CLI
$ npx skills add WrongStack/WrongStack --skill web-platform-baseline -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install WrongStack/WrongStack web-platform-baseline --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/WrongStack/WrongStack.git skills-src && mkdir -p .claude/skills && cp -r skills-src/packages/core/skills/web-platform-baseline .claude/skills/web-platform-baseline && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
web-platform-baseline
GitHub stars
368
Token cost
~1.6k tokens
SKILL.md length
791 words
Files
4 (incl. references)
Skills in repo
38
Repo updated
First seen
Licence
MIT

At a glance

Use this skill before asserting that a CSS, HTML or accessibility capability is available, unavailable, or the right tool — it carries dated, refreshable platform facts and refuses to let stale…

  • Works in 2 steps: Stale caution — writing a media-query… → Stale confidence — asserting support for…
  • Tasks that involve Positioning and messaging
  • SKILL.md covers Why this exists, The staleness rule —…, What's in here and Rules, plus 2 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Web Platform Baseline is an agent skill from WrongStack/WrongStack. Use this skill before asserting that a CSS, HTML or accessibility capability is available, unavailable, or the right tool — it carries dated, refreshable platform facts and refuses to let stale knowledge be stated as current. Triggers: user says "container query", "can I use", "browser support", "modern CSS", ":has", "subgrid", "view transition", "popover", "dialog", "anchor positioning", "scroll-driven animation", "baseline", "polyfill", "is it supported", "fallback needed".

Its SKILL.md is about 1.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 4 other files, including reference files (for example `references/a11y.md`, `references/css.md` and `references/html.md`).

It sits in Frontend & Design, covering Positioning and messaging. The repository describes itself as: An AI coding agent that reads your code, edits files, runs commands, and reasons through bugs — across a terminal REPL, a full-screen TUI, and a browser UI, while you keep your… The licence is MIT.

When your agent uses it

  • Tasks that involve Positioning and messaging

Example prompts

  • “container query”
  • “can I use”
  • “browser support”
  • “/web-platform-baseline”

Workflow steps

2 steps, taken from the first numbered list in SKILL.md.

  1. Stale caution — writing a media-query ladder, a JS resize observer, or a
  2. Stale confidence — asserting support for something that is still behind

What it can do on your machine

Read from SKILL.md and the folder at commit ec76a20. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Web Platform Baseline loads about 1.6k tokens when it runs, and up to ~5.2k if it reads all its reference files. Until then it costs about 126 tokens; SKILL.md has 791 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~126
When it runs · the whole SKILL.md, loaded when a task matches
~1.6k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~5.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from WrongStack/WrongStack at commit ec76a20, republished under its MIT licence (© WrongStack). 791 words, ~1,587 tokens.

Download SKILL.mdSave it as .claude/skills/web-platform-baseline/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.
name
web-platform-baseline
description
Use this skill before asserting that a CSS, HTML or accessibility capability is available, unavailable, or the right tool — it carries dated, refreshable platform facts and refuses to let stale knowledge be stated as current. Triggers: user says "container query", "can I use", "browser support", "modern CSS", ":has", "subgrid", "view transition", "popover", "dialog", "anchor positioning", "scroll-driven animation", "baseline", "polyfill", "is it supported", "fallback needed".
version
1.0.0
required-capabilities
filesystem.read
required-tools
skill, search, fetch
optional-capabilities
web.research

Web Platform Baseline — WrongStack

Why this exists

Two failure modes, both expensive:

  1. Stale caution — writing a media-query ladder, a JS resize observer, or a z-index hack for something the platform has shipped for years. This is the most common source of dated-looking, over-engineered UI code.
  2. Stale confidence — asserting support for something that is still behind a flag, or naming a syntax that changed during standardization.

Model training data guarantees both. This skill is the correction: every fact lives in a dated reference file, and a fact past its shelf life may not be asserted without a live check.

The staleness rule — non-negotiable

Every reference file in this skill starts with a verified: date.

<!-- verified: 2026-09-16 | sources: MDN, web.dev Baseline, caniuse -->

Before you assert any fact from a reference:

Age of verified:What you may do
≤ 90 daysState it as current. Cite the file.
> 90 daysState it as "as of <date>", and verify with search / fetch before it drives an architectural decision.
> 180 days, or the fact decides the architectureVerify first, then answer. Update the reference file with the new date and the corrected fact.

Never silently "refresh" a fact from memory. A date you cannot evidence is a guess wearing a date.

This has already happened once, in this file's own references. The first revision of css.md and html.md was written from model knowledge and stamped with that day's date without a single live check. A later verification pass found 7 of 14 tiers wrong — anchor positioning, text-box-trim and field-sizing were marked unavailable months after they had reached Baseline (the "stale caution" failure, which makes code hand-roll what the platform already does), while scroll-driven animations, text-wrap: pretty and the Popover API were marked more available than they are (the "stale confidence" failure). So: when you write a verified: date, the check must have happened in that turn, and the sources must be named on the line. If you cannot verify a row, mark it UNVERIFIED and say so rather than guessing a tier. The same rule applies to anything you remember about framework defaults (Tailwind, React, the kit stacks) — those live in tech-stack and research-web, and rot the same way.

When you do verify, use the primary sources in this order: MDN (syntax and semantics), web.dev Baseline (availability tier), caniuse (version detail), the spec (when behavior is genuinely contested). Two agreeing sources minimum, per the research-web rules.

What's in here

Load on demand — do not paste these into a plan wholesale:

Load these with the skill tool — skill({ name: "web-platform-baseline", resource: "references/<file>.md" }):

  • references/css.md — layout, container queries, :has(), subgrid, nesting, color-mix(), OKLCH, text-wrap, cascade layers, scroll-driven animation, @starting-style, anchor positioning
  • references/html.md — <dialog>, popover, <details> interop, form controls, field-sizing, lazy/priority hints, view transitions
  • references/a11y.md — WCAG 2.2 additions, focus appearance, target size, prefers-* queries, accessible names, live regions
Show full SKILL.md (328 more words)Show less

Rules

  1. Never assert browser support from memory. Cite a reference file (with its date) or verify live.
  2. Prefer the platform. If CSS or HTML does it natively, a JS dependency is a regression. Reach for a library only when the native path is genuinely missing a capability the product needs — and say which one.
  3. Progressive enhancement over polyfill. @supports, a sane fallback, and the modern path on top. Shipping a polyfill for a Baseline-available feature is dead weight.
  4. Availability tier, not a version list. Say "Baseline widely available" or "Baseline newly available (needs a fallback)" — version tables age badly and users rarely need them.
  5. When the answer is "it depends on your support target", ask. Do not assume evergreen-only, and do not assume enterprise IE-era constraints.
  6. Correct the file, not just the answer. If you verify a fact and find the reference stale, update it and its verified: date in the same turn. A correction you don't write down will be re-derived from scratch next week.

Applying it

The common wins, which the references cover in detail:

Old reflexCurrent platform
JS resize observer to restyle a componentcontainer queries
Parent-state class toggling in JS:has()
Nested grid hacks for alignment across cardssubgrid
Manual focus-trap library for a modal<dialog> + showModal()
Positioned dropdown with a JS collision librarypopover + anchor positioning
scroll listener driving an animationscroll-driven animations
SASS just for nesting and variablesnative nesting + custom properties
Ragged headline hand-fixed with <br>text-wrap: balance
Hand-mixed hover/disabled color variantscolor-mix() on the kit token

Each of these makes the code both shorter and less generated-looking — the hand-rolled version of a platform feature is itself a tell.

Skills in scope

  • design-craft — the composition rules these capabilities serve
  • design-system — kit tokens; platform features consume tokens, never literals
  • design-critique — auditing an existing UI against current practice
  • research-web — the verification workflow when a fact is stale
  • tech-stack — framework/package version currency (the same rot, different layer)

© WrongStack, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 3 other files (references) in packages/core/skills/web-platform-baseline of WrongStack/WrongStack.

  • SKILL.md
  • references/a11y.md
  • references/css.md
  • references/html.md

Open the folder on GitHubat commit ec76a20

Compare with similar skills

Web Platform Baseline next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Web Platform Baseline compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Web Platform Baseline this skillWrongStack/WrongStack368—~1.6kAutomated safety check: PassMIT
Landing Page Copywriterjulianromli/ai-skills1912 repos~1.4kAutomated safety check: PassNone
Ads Competitorszubair-trabzada/ai-ads-claude267—~4.9kAutomated safety check: PassMIT
Brand SystemsHermeticOrmus/LibreUIUX-Claude-Code111—~2.3kAutomated safety check: PassMIT
Base UI Reactsecondsky/claude-skills227—~1.9kAutomated safety check: PassMIT
Design Archivistcuriositech/some_claude_skills2431 repos~1.6kAutomated safety check: PassMIT

Similar skills

  • Landing Page Copywriter

    julianromli/ai-skills

    Write high-converting landing page copy using proven frameworks like PAS (Problem-Agitate-Solution), AIDA, and StoryBrand.

    191 GitHub starsUsed in 2 repos~1.4k tokens
    Frontend & DesignAuto-check passed
  • Ads Competitors

    zubair-trabzada/ai-ads-claude

    Competitive Ad Intelligence. An agent skill from zubair-trabzada/ai-ads-claude.

    267 GitHub stars~4.9k tokensUpdated 6 mo ago
    Frontend & DesignAuto-check passed
  • Brand Systems

    HermeticOrmus/LibreUIUX-Claude-Code

    Brand identity building blocks: positioning, personality, archetypes, logo systems, color palettes, type pairing, voice, and guidelines documentation.

    111 GitHub stars~2.3k tokensUpdated 2 days ago
    Frontend & DesignAuto-check passed
  • Base UI React

    secondsky/claude-skills

    MUI Base UI unstyled React components with Floating UI. An agent skill from secondsky/claude-skills.

    227 GitHub stars~1.9k tokensUpdated 9 days ago
    Frontend & DesignAuto-check passed
  • Design Archivist

    curiositech/some_claude_skills

    Long-running design anthropologist that builds comprehensive visual databases from 500-1000 real-world examples, extracting color palettes, typography patterns, layout systems, and interaction…

    243 GitHub starsUsed in 1 repo~1.6k tokens
    Frontend & DesignAuto-check passed
  • Product Appeal Analyzer

    curiositech/some_claude_skills

    Evaluate product desirability, market positioning, and emotional resonance—the complement to friction analysis.

    243 GitHub starsUsed in 1 repo~2.4k tokens
    Frontend & DesignAuto-check passed

More from WrongStack/WrongStack

All 38 skills in this repo
  • Design Craft

    WrongStack/WrongStack

    Design or substantially improve user-facing interfaces with a product-specific visual direction, content hierarchy, and rendered critique.

    368 GitHub stars~2k tokensUpdated today
    Auto-check passed
  • Design Critique

    WrongStack/WrongStack

    A skill your agent uses to audit an interface that already exists and say precisely why it looks generated, templated, or unfinished — a scored rubric across composition, typography, color, states…

    368 GitHub stars~3k tokensUpdated today
    Auto-check passed
  • Mailbox Bridge

    WrongStack/WrongStack

    A skill your agent uses when external coding agents (Claude Code, Aider, custom scripts) need to participate in the project's shared WrongStack mailbox, or when a user asks to "expose the mailbox"…

    368 GitHub stars~3.9k tokensUpdated today
    Auto-check passed
  • Multi Agent

    WrongStack/WrongStack

    A skill your agent uses whenever work can be split across multiple AI agents running in parallel, or when orchestrating leader/worker patterns in WrongStack.

    368 GitHub stars~3.6k tokensUpdated today
    Auto-check passed
  • Wrongstack Mailbox

    WrongStack/WrongStack

    A skill your agent uses when the user wants to communicate with WrongStack's shared project mailbox from outside WrongStack — read messages sent by WrongStack agents, send replies, broadcast to all…

    368 GitHub stars~3.5k tokensUpdated today
    Auto-check passed
  • API Design

    WrongStack/WrongStack

    A skill your agent uses when designing, implementing, or reviewing an HTTP API — endpoints, request and response shapes, errors, pagination, versioning, and authorization.

    368 GitHub stars~1.3k tokensUpdated today
    Auto-check passed

Questions about Web Platform Baseline

What does Web Platform Baseline do?

Use this skill before asserting that a CSS, HTML or accessibility capability is available, unavailable, or the right tool — it carries dated, refreshable platform facts and refuses to let stale…. Web Platform Baseline is an agent skill from WrongStack/WrongStack. Use this skill before asserting that a CSS, HTML or accessibility capability is available, unavailable, or the right tool — it carries dated, refreshable platform facts and refuses to let stale knowledge be stated as current.

When should I use Web Platform Baseline?

Web Platform Baseline fits situations like: tasks that involve Positioning and messaging.

How do I install Web Platform Baseline in Claude Code?

Run `npx skills add WrongStack/WrongStack --skill web-platform-baseline -a claude-code`. Or copy the skill folder (packages/core/skills/web-platform-baseline in WrongStack/WrongStack) into .claude/skills/web-platform-baseline in your project. Claude Code loads it when a task matches its description.

How do I install Web Platform Baseline in Codex?

Run `npx skills add WrongStack/WrongStack --skill web-platform-baseline -a codex`. Or copy the skill folder (packages/core/skills/web-platform-baseline in WrongStack/WrongStack) into .agents/skills/web-platform-baseline in your project. Codex loads it when a task matches its description.

Can I use Web Platform Baseline in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add WrongStack/WrongStack --skill web-platform-baseline -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/web-platform-baseline, .gemini/skills/web-platform-baseline, .github/skills/web-platform-baseline and .opencode/skills/web-platform-baseline in your project.

What does Web Platform Baseline need to run?

SKILL.md names no scripts, command-line tools or credentials: Web Platform Baseline is instructions for the agent only.

Does Web Platform Baseline access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Web Platform Baseline safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Web Platform Baseline use?

Web Platform Baseline is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Web Platform Baseline use?

About 1.6k tokens (SKILL.md is roughly 6.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 3.6k tokens, read only when the agent opens those files.

What are the alternatives to Web Platform Baseline?

Skills that share tags, products or a category with Web Platform Baseline: Landing Page Copywriter (julianromli/ai-skills, 191 stars), Ads Competitors (zubair-trabzada/ai-ads-claude, 267 stars), Brand Systems (HermeticOrmus/LibreUIUX-Claude-Code, 111 stars) and Base UI React (secondsky/claude-skills, 227 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Web Platform Baseline?

WrongStack (a GitHub organization) maintains it in WrongStack/WrongStack, which has 368 GitHub stars. The repository holds 38 skills in this directory. The repository was last updated on October 6, 2026.

Source: WrongStack/WrongStack on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.