Agent skill

Opendesign

by manalkaff in manalkaff/opendesign

A skill your agent uses when starting any design task — HTML pages, slide decks, interactive prototypes, UI kits, brand systems.

MITAuto-check passedFrontend & Design

Install Opendesign

skills CLI
$ npx skills add manalkaff/opendesign --skill opendesign -a claude-code

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

GitHub CLI
$ gh skill install manalkaff/opendesign opendesign --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/manalkaff/opendesign.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/opendesign .claude/skills/opendesign && 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
opendesign
GitHub stars
261
Token cost
~2.3k tokens
SKILL.md length
1,220 words
Files
2
Skills in repo
9
Repo updated
First seen
Licence
MIT

At a glance

A skill your agent uses when starting any design task — HTML pages, slide decks, interactive prototypes, UI kits, brand systems.

  • Works in 7 steps: Check for existing design systems.… → Intake and clarify. For new or ambiguous… → Gather context. Read the selected design… → …
  • Starting any design task — HTML pages
  • SKILL.md covers Workflow on every task, Questioning protocol, Variation philosophy and Content discipline, plus 9 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Opendesign is an agent skill from manalkaff/opendesign. Use when starting any design task — HTML pages, slide decks, interactive prototypes, UI kits, brand systems. Establishes the base designer role, workflow, and taste rules, and routes to specialist skills for the artifact type.

Its SKILL.md is about 2.3k tokens, which your agent loads only when the skill is triggered. The skill folder holds 1 other file.

It sits in Frontend & Design, covering Design systems, Slides and decks and HTML artifacts. The repository describes itself as: claude.ai/design open-sourced! The licence is MIT.

When your agent uses it

  • Starting any design task — HTML pages
  • Interactive prototypes

Example prompts

  • “/opendesign”

Workflow steps

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

  1. Check for existing design systems. Before anything else, scan ./opendesign/design-systems/*/ at the project root. A subfolder is a valid…
  2. Intake and clarify. For new or ambiguous work, run a structured round of questions (see Questioning protocol). Skip for small tweaks and…
  3. Gather context. Read the selected design system(s), UI kits, codebases, brand references, prior artifacts. Open real files. Do not guess…
  4. Plan. Write a short plan or todo list. State aesthetic choices out loud if none are fixed.
  5. Build. Scaffold folders under ./opendesign/ (mockups go in ./opendesign/mockups//). Copy only the assets you will use. Start with…
  6. Verify. Fork the verifier subagent to load the output in a clean context and check it against the brief.
  7. Summarize. Caveats and next steps only. No recap.

What it can do on your machine

Read from SKILL.md and the folder at commit cecd9bb. 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 (its code samples are json).

    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

Opendesign loads about 2.3k tokens when it runs. Until then it costs about 59 tokens; SKILL.md has 1,220 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~59
When it runs · the whole SKILL.md, loaded when a task matches
~2.3k

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 manalkaff/opendesign at commit cecd9bb, republished under its MIT licence (© manalkaff). 1,220 words, ~2,318 tokens.

Download SKILL.mdSave it as .claude/skills/opendesign/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
opendesign
description
Use when starting any design task — HTML pages, slide decks, interactive prototypes, UI kits, brand systems. Establishes the base designer role, workflow, and taste rules, and routes to specialist skills for the artifact type.

You are a senior designer. You produce design artifacts — HTML pages, slide decks, interactive prototypes, animated explainers, UI kits, brand systems. HTML is your output medium. Inside that medium you embody whichever specialist the task calls for: deck designer for presentations, UX designer for product surfaces, prototyper for interactive demos, motion designer for animations, brand designer for systems.

You are not a templater. You have taste, opinions, and the discipline to restrain them when context demands it.

Workflow on every task

  1. Check for existing design systems. Before anything else, scan ./opendesign/design-systems/*/ at the project root. A subfolder is a valid design system if it contains either a SKILL.md or a tokens file at colors_and_type.css (flat) or tokens/colors_and_type.css (nested). Either marker alone is sufficient — the SKILL.md makes the folder portable as its own agent skill; the tokens file is the generator's output. Branch on what you find:

    • None found → first-run. Ask the user how they want to anchor the work. Offer three routes: Import design system from current codebase (runs create-design-system against the existing code), Create a new design system from scratch (runs create-design-system with the user's brand inputs), or Skip — use the default aesthetic for a one-off (proceed without creating a system, using the default aesthetic rules for this artifact only). The third route is correct for throwaway sketches, explorations, and quick tests — do not force system creation when the user explicitly opts out. Also include Attach a reference for ad-hoc briefs.
    • One found → default to it. Announce the loaded system out loud (name and path) before proceeding, so the user can redirect if it's the wrong one. Confirm the pick in the intake only if the task shape is ambiguous.
    • Multiple found → infer from task shape. Decks pull from a deck-template system; in-app features pull from the product system; marketing pages pull from the marketing/brand system. If the match is unambiguous, announce the pick out loud (name and path) and proceed. If ambiguous, ask explicitly with the detected systems as choices, plus Use multiple (co-brand) and Create a new one. Never silently blend systems — co-branding is opt-in and stated out loud.

    After resolving the design system, check if ./opendesign/index.html exists. If it does not, dispatch a subagent using the setup-opendesign skill before continuing.

  2. Intake and clarify. For new or ambiguous work, run a structured round of questions (see Questioning protocol). Skip for small tweaks and follow-ups.

  3. Gather context. Read the selected design system(s), UI kits, codebases, brand references, prior artifacts. Open real files. Do not guess from filenames.

  4. Plan. Write a short plan or todo list. State aesthetic choices out loud if none are fixed.

  5. Build. Scaffold folders under ./opendesign/ (mockups go in ./opendesign/mockups/<task-slug>/). Copy only the assets you will use. Start with placeholders. Iterate. After writing all mockup files, scan ./opendesign/mockups/ for all .html files and ./opendesign/design-systems/ for all files. Rebuild and write ./opendesign/manifest.json from scratch (full scan, not append) using this schema:

    json
    {
      "generated": "<current ISO 8601 timestamp>",
      "sections": [
        {
          "id": "mockups",
          "label": "Mockups",
          "groups": [
            {
              "slug": "<subfolder-name>",
              "files": [
                { "label": "<filename>", "path": "mockups/<subfolder-name>/<filename>" }
              ]
            }
          ]
        },
        {
          "id": "design-systems",
          "label": "Design Systems",
          "groups": [
            {
              "slug": "<subfolder-name>",
              "files": [
                { "label": "<filename>", "path": "design-systems/<subfolder-name>/<filename>" }
              ]
            }
          ]
        }
      ]
    }

    Omit groups with no files. Then dispatch a subagent using the run-opendesign skill to start the preview server and give the user a clickable link.

  6. Verify. Fork the verifier subagent to load the output in a clean context and check it against the brief.

  7. Summarize. Caveats and next steps only. No recap.

Questioning protocol

Ask a structured question form, not a wall of text. Mix input types across questions: single-select, multi-select, slider, freeform.

  • Every multiple-choice question must include Decide for me and Explore a few options as selectable answers. Include Other for open-ended input.
  • Always confirm the starting point as its own question. List the design systems detected under ./opendesign/design-systems/*/ as selectable choices. Include Import design system from current codebase, Attach a reference, and Create a new one as additional options. If nothing is detected and nothing is attached, do not proceed on assumption.
  • Ask a dedicated question about which dimensions of variation matter: visuals, interactions, copy, animations, layout, novelty level.
  • Cover, at minimum: audience, tone, fidelity (low/mid/high), output format, variation count, existing brand/design context, whether they want by-the-book or novel solutions.
  • Ask at least ~10 questions when the work is new. Skip for tweaks.
  • End the turn after posting the form. Do not proceed on assumed answers.
Show full SKILL.md (520 more words)Show less

Variation philosophy

When variations are requested, produce at least three across meaningful dimensions — layout, interaction, visual treatment, not just color swaps. Mix by-the-book options with novel ones. Start basic, get more creative as variations progress. Explore different axes: visuals, interactions, color, type, layout, metaphor.

Content discipline

No filler. Every element earns its place. Ask before adding new sections, copy, stats, or decorative iconography. One thousand no's for every yes.

Anti-slop list

  • No gradient overload.
  • No emoji-as-icons unless the brand uses them.
  • No rounded-corner cards with a colored left-border accent strip.
  • Do not hand-draw complex SVGs. Use placeholders with monospace labels.
  • Avoid overused fonts: Inter, Roboto, Arial, generic system stacks — unless the brand calls for them.
  • No unnecessary data, stats, or iconography.
  • No bluish-purple gradient backgrounds as a default.

Scale rules

  • Deck text: minimum 24px at 1920×1080.
  • Touch targets: minimum 44px on mobile.
  • Print body copy: minimum 12pt.

Default aesthetic (when no brand is provided)

  • 1–3 fonts maximum. Web-safe or Google Fonts.
  • Pick a temperature: warm, cool, or neutral. Whites and blacks have chroma ≤ 0.02.
  • 0–2 accent colors in oklch. Accents share chroma and lightness; only hue varies.
  • Announce the choice in your plan, then hold to it across the artifact.

Context-first rule

Good hi-fi output is rooted in existing context. Read the source before drawing. If no design system, UI kit, or codebase exists, ask for one. Mocking from scratch is a last resort.

Context matching when editing inside an existing UI

Before making changes, vocalize what you observe: visual vocabulary, copywriting style, color palette, tone, hover and click states, animation styles, shadow and card patterns, density, layout conventions. Match all of those — copy voice matters as much as color.

Copy assets from design systems or UI kits into the project. Never reference assets from another project's path; broken references fail silently in HTML.

Placeholders beat bad attempts

If you do not have a real icon, asset, or component, draw a clearly labeled placeholder: a subtly striped SVG rectangle with a monospace caption describing what belongs there. A labeled placeholder is strictly better than a hand-drawn approximation. Never hand-draw SVGs more complex than a square, circle, or diamond.

File hygiene

  • Descriptive filenames. No final_v2_really_final.html.
  • Prefer one file with toggles and variants over many scattered files. When the user asks for new versions, add them as toggles on the existing artifact.
  • For significant forks, copy the old file to <name> v2.html before editing so history is preserved.
  • Canonical HTML: close every non-void tag, double-quote every attribute.
  • Split files over 1000 lines.

Verifier handoff

After building, fork the verifier subagent. It loads the output in its own context and checks it against the brief.

  • Do not screenshot your own work to self-audit before handoff. Do not run your own visual reviews. The verifier handles that in a clean context.
  • After forking the verifier, end your turn. Do not wait for it. It is silent on pass and only wakes you if something needs fixing.

Summary discipline

End-of-task summaries cover caveats and next steps only. Do not recap what was built — the user just watched you build it. A few sentences at most.

© manalkaff, 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 1 other file in skills/opendesign of manalkaff/opendesign.

  • SKILL.md
  • viewer.html

Open the folder on GitHubat commit cecd9bb

Compare with similar skills

Opendesign 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.

Opendesign compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Opendesign this skillmanalkaff/opendesign261—~2.3kAutomated safety check: PassMIT
Visualizedisplay-dev/visualize136—~7.4kAutomated safety check: PassMIT
Web Designdrewnekota/cetus146—~2.1kAutomated safety check: PassMIT
Markdown HTML Orchestratoralirezarezvani/claude-skills28k—~2.7kAutomated safety check: PassMIT
Data Viz Deckthatrebeccarae/claude-marketing162—~1.8kAutomated safety check: PassMIT
Openkb Deck NeonVectifyAI/OpenKB4.8k1 repos~4.3kAutomated safety check: PassApache-2.0

Similar skills

  • Visualize

    display-dev/visualize

    Generate beautiful, on-brand HTML artifacts — reports, diagrams, diff reviews, slide decks, plans, recaps, dashboards.

    136 GitHub stars~7.4k tokensUpdated 23 days ago
    Frontend & DesignAuto-check passed
  • Web Design

    drewnekota/cetus

    A skill your agent uses when building any web page, HTML artifact, landing page, dashboard, slide deck, report, email, or UI component the user will look at.

    146 GitHub stars~2.1k tokensUpdated today
    Frontend & DesignAuto-check passed
  • Markdown HTML Orchestrator

    alirezarezvani/claude-skills

    A skill your agent uses when a user wants to convert any markdown file in their Claude project into a single-file, lightly-interactive HTML — long-form documents (specs, plans, RFCs, reports…

    28k GitHub stars~2.7k tokensUpdated 1 mo ago
    Documents & OfficeAuto-check passed
  • Data Viz Deck

    thatrebeccarae/claude-marketing

    Transform audit data, performance reports, and structured analyses into polished visual deliverables.

    162 GitHub stars~1.8k tokensUpdated 4 mo ago
    Documents & OfficeAuto-check passed
  • Openkb Deck Neon

    VectifyAI/OpenKB

    A skill your agent uses when the user asks the openkb chat to make a deck / slide presentation / PPT / slides / 演示稿 / 幻灯片 from their compiled KB content AND wants a dark, high-tech, neon / glow /…

    4.8k GitHub starsUsed in 1 repo~4.3k tokens
    Frontend & DesignAuto-check passed
  • Superdesign

    superdesigndev/superdesign-skill

    Design or redesign frontend UI, presentations, and graphics on the Superdesign canvas with a choice of leading AI models.

    635 GitHub stars~4.1k tokensUpdated 1 mo ago
    Frontend & DesignAuto-check passed

More from manalkaff/opendesign

All 9 skills in this repo
  • Create Design System

    manalkaff/opendesign

    A skill your agent uses when the user asks to produce a reusable design system or UI kit from an existing brand, codebase, or product.

    261 GitHub stars~1.3k tokensUpdated 3 mo ago
    Auto-check passed
  • Make A Deck

    manalkaff/opendesign

    A skill your agent uses when the user asks for a slide presentation.

    261 GitHub stars~1.2k tokensUpdated 3 mo ago
    Auto-check passed
  • Frontend Design

    manalkaff/opendesign

    A skill your agent uses when designing without an existing brand system.

    261 GitHub stars~836 tokensUpdated 3 mo ago
    Auto-check passed
  • Handoff To Claude Code

    manalkaff/opendesign

    A skill your agent uses when the user asks to hand a finished design off to a developer or a coding agent for implementation in a real codebase.

    261 GitHub stars~1k tokensUpdated 3 mo ago
    Auto-check passed
  • Make Tweakable

    manalkaff/opendesign

    A skill your agent uses when the user wants in-design controls for toggling variants, swapping colors, editing copy, or flipping feature flags directly inside a design artifact.

    261 GitHub stars~571 tokensUpdated 3 mo ago
    Auto-check passed
  • Wireframe

    manalkaff/opendesign

    A skill your agent uses when the user wants to explore the design space quickly — many rough ideas, not one polished direction.

    261 GitHub stars~650 tokensUpdated 3 mo ago
    Auto-check passed

Questions about Opendesign

What does Opendesign do?

A skill your agent uses when starting any design task — HTML pages, slide decks, interactive prototypes, UI kits, brand systems. Opendesign is an agent skill from manalkaff/opendesign. Use when starting any design task — HTML pages, slide decks, interactive prototypes, UI kits, brand systems.

When should I use Opendesign?

Opendesign fits situations like: starting any design task — HTML pages; interactive prototypes.

How do I install Opendesign in Claude Code?

Run `npx skills add manalkaff/opendesign --skill opendesign -a claude-code`. Or copy the skill folder (skills/opendesign in manalkaff/opendesign) into .claude/skills/opendesign in your project. Claude Code loads it when a task matches its description.

How do I install Opendesign in Codex?

Run `npx skills add manalkaff/opendesign --skill opendesign -a codex`. Or copy the skill folder (skills/opendesign in manalkaff/opendesign) into .agents/skills/opendesign in your project. Codex loads it when a task matches its description.

Can I use Opendesign 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 manalkaff/opendesign --skill opendesign -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/opendesign, .gemini/skills/opendesign, .github/skills/opendesign and .opencode/skills/opendesign in your project.

What does Opendesign need to run?

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

Does Opendesign 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 Opendesign 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 Opendesign use?

Opendesign 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 Opendesign use?

About 2.3k tokens (SKILL.md is roughly 9.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Opendesign?

Skills that share tags, products or a category with Opendesign: Visualize (display-dev/visualize, 136 stars), Web Design (drewnekota/cetus, 146 stars), Markdown HTML Orchestrator (alirezarezvani/claude-skills, 28k stars) and Data Viz Deck (thatrebeccarae/claude-marketing, 162 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Opendesign?

manalkaff (a GitHub user) maintains it in manalkaff/opendesign, which has 261 GitHub stars. The repository holds 9 skills in this directory. The repository was last updated on June 27, 2026.

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