agentleFS
atlassianatlassian/atlassian-mcp-server/skills/atlassian-design/SKILL.md

atlassian-design

Skill1.2k starsChanged 3 months agoApache-2.0View on GitHub

Apply the Atlassian Design System (ADS) to artifacts, pages, mockups and prototypes — official tokens (light and dark), type, spacing, elevation, icons, component recipes, smart links, content and accessibility rules. Use when asked for Atlassian-style or ADS-styled UI.

2 other projects use a version of this.

Trending

More talk lately.about the whole repository

Dec 2025: 1 mentionJan 2026: 1 mentionFeb 2026: 1 mentionMar 2026: 1 mentionApr 2026: 5 mentionsMay 2026: 3 mentionsJun 2026: 1 mentionJul 2026: 1 mentionAug 2026: 3 mentionsSep 2026: 6 mentionsOct 2026: 15 mentionsNov 2026: 4 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the atlassian/atlassian-mcp-server repository that our weekly crawl found. There is no adoption data for this file yet.

Most recently taken up by

Who's talking about atlassian/atlassian-mcp-server

The biggest 5 of 44 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 44

Talked about on GitHub and Bluesky (44 times).

On GitHub · 43

38 more on GitHub

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to make an artifact look and read like an Atlassian app (Jira, Confluence, Loom, Atlas style): pages, dashboards, …

    line 8
  2. Distinctive:

    Support light and dark: Every token has a dark value.

    line 32
  3. Distinctive:

    Case and punctuation: Use sentence case for all UI text.

    line 40
  4. Distinctive:

    Contractions: Use contractions and curly apostrophes (’).

    line 41
  5. Distinctive:

    Abbreviations: Avoid "e.g.", "i.e.", "etc." and "&".

    line 42
  6. Distinctive:

    Empty states: Use a short title, one or two sentences on why it's empty and what's next, and one primary CTA with an imperative verb …

    line 46
  7. Distinctive:

    Language: Use inclusive language. Use "they" when a person's pronouns aren't known.

    line 47
  8. Very specific:

    Dates and times: Use relative times for recent events ("5 minutes ago").

    line 48

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Atlassian Design System (ADS) for artifacts

Use this skill to make an artifact look and read like an Atlassian app (Jira, Confluence, Loom, Atlas style): pages, dashboards, prototypes, mockups, trackers and reports.

This skill is distilled from the official ADS MCP server (https://mcp.atlassian.com/v1/ads/public/mcp) and the @atlaskit/tokens, @atlaskit/button, @atlaskit/lozenge, @atlaskit/smart-card and @atlaskit/icon packages. Token values match @atlaskit/tokens 20.x.

How this fits with other skills

  • The page contract. If the session has an artifact skill or quickstart, follow its page contract, and use this skill for the palette, typography and components. Where they conflict on look and feel, ADS wins. Where they conflict on the contract, the contract wins. Without one, use this contract:
    • one self-contained HTML file, with all CSS, JavaScript, icons and images inline, and no CDN or other external requests
    • a meaningful <title> and <meta name="viewport" content="width=device-width, initial-scale=1">
    • a 16px gutter at phone width, and no horizontal page scroll
    • light values on :root, dark values under :root[data-theme="dark"] and under @media (prefers-color-scheme: dark) guarded by :root:not([data-theme="light"]) (see render_kit/tokens.css)
    • no localStorage, sessionStorage or IndexedDB
    • a small file: inline only the icons you use, and don't embed raw data you don't display
    • for a live artifact whose host supports connectors (see teamwork-graph-context), data comes only through the host's connectors and storage, never by fetching URLs, and browser storage is still off
  • When to load this skill. For an answer-first skill, use ordinary Markdown and source links in the initial response. Load this design guidance only after the user requests an HTML report or other visual artifact. Other skills keep their existing artifact flow.
  • The page kit. Skills with a render.py build their page through the kit in render_kit/ in this folder: the agent writes a JSON data file, and the renderer builds the page from the kit's components, with the ADS tokens in light and dark, smart links, icons, tabs and copy buttons. Their Build step says when to read references/data-format.md. To build a page by hand, follow references/hand-built.md.
  • For charts, use a charting skill for form and interaction only if one is available in the session. Otherwise use the data visualization rules in references/hand-built.md. In either case, use ADS chart tokens and palette; do not assume an extra skill is installed.
  • If the Atlassian Design System connector (tools named like ads_plan, ads_search_tokens, ads_search_components, ads_search_icons, ads_get_guidelines, ads_get_a11y_guidelines) is enabled in the session, call ads_plan first to look up tokens, icons and components for the task, and ads_get_guidelines for foundations and content questions. The connector is more current than this file.
  • Without the connector, this skill and its references are enough for most artifacts. For anything it doesn't cover, fetch the public docs at https://atlassian.design (for example /components/<name>/usage, /foundations/color, /components/tokens/all-tokens).

Non-negotiables

  1. Never hard-code a color, space, radius, shadow or font value. Use the --ds-* custom properties in render_kit/tokens.css. If you need a value it doesn't have, pull it from the npm package (see "Getting any other token" in references/hand-built.md).
  2. Choose tokens by role, not by how they look. Brand is for primary actions and selection. Danger, warning, success, information and discovery carry meaning. Accents (blue, teal, green, lime, yellow, orange, red, magenta, purple, gray) are for categories and user-chosen colors only.
  3. Support light and dark. Every token has a dark value. Ship both.
  4. Use sentence case everywhere: titles, headings, buttons, tabs, labels and menu items. Use all caps only for acronyms and the legacy lozenge.
  5. Meet WCAG AA. Text under 24px needs 4.5:1 contrast. Large text and essential UI need 3:1. Never use color alone to carry meaning.
  6. Every link to a page, work item or other object is an inline smart link. Never show a bare URL, a plain blue text link or "click here". A renderer builds them from the link data in references/data-format.md. By hand, follow "Smart links" in references/hand-built.md.

Content (voice and writing)

  • Voice. Bold, optimistic, practical with a wink. Be direct and concise, and put the important bits first. Know when to be serious.
  • Case and punctuation. Use sentence case for all UI text. No periods in headings, titles, buttons or list fragments. Rephrase question headings as statements. Avoid gerunds in UI headings ("Create a space," not "Creating a space").
  • Contractions. Use contractions and curly apostrophes (’).
  • Abbreviations. Avoid "e.g.", "i.e.", "etc." and "&". Spell out product and feature names.
  • Errors. Say what happened and what to do next. Use "we," don't blame the person ("We couldn't save your changes," not "You entered invalid data").
  • Warnings. State the consequence before the action.
  • Success. Confirm the result, then get out of the way.
  • Empty states. Use a short title, one or two sentences on why it's empty and what's next, and one primary CTA with an imperative verb ("Create," "Try"). The tone is motivating for a blank slate, celebratory for all-done and neutral for no results.
  • Language. Use inclusive language. Use "they" when a person's pronouns aren't known.
  • Dates and times. Use relative times for recent events ("5 minutes ago"). Use unambiguous dates such as "Sep 30, 2026" rather than 09/30/26.

Where to go next

  • Writing the data for a skill's renderer: references/data-format.md. The skill's own Build step and example.json give its keys.
  • Building a page by hand, or when Python can't run: references/hand-built.md. It has the foundations, the component recipes, the smart link rules, live states, the accessibility checklist and the checks before publishing.
  • Building or changing a renderer: references/components.md, every component in the kit, generated from its docstrings. New visual pieces go into the kit as components, not into a skill's render.py.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.