Mstar Design Md logo

Mstar Design Md

Community
btspoony
mstar-design-md

DESIGN.md 设计系统规范:创建/审计/维护 design tokens、三级完整性清单、light/dark 双主题(DESIGN.dark.md)。用户提到 DESIGN.md / design tokens / design system,或实现、审查 UI 时读取。

Overview

Publisherbtspoony
Repositorymstar-harness
Skill namemstar-design-md
Stars
67
Forks
4
Bundled files
5
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 5 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by btspoony on GitHub. Read the source before you install it.

Installation

Install the Mstar Design Md AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/btspoony/mstar-harness/tree/main/skills/mstar-design-md
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/btspoony/mstar-harness.git /tmp/mstar-harness
mkdir -p .claude/skills
cp -r /tmp/mstar-harness/skills/mstar-design-md .claude/skills/mstar-design-md
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Mstar Design Md in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Mstar Design Md on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Mstar Design Md is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Load order

Before first Read of this skill: Read mstar-harness-core (SKILL.md). For Prepare phase integration and gate rules, read mstar-phase-gates. For plan directory paths ({HARNESS_DIR}, {SPECS_DIR}), read mstar-conventions. On conflict, mstar-harness-core wins.

你还可能要 Read何时
mstar-phase-gatesPrepare 阶段判定 gate、何时 DESIGN.md 必须就绪
mstar-conventions{HARNESS_DIR} / {SPECS_DIR} 路径解析
mstar-roles@architect / @product-manager / @frontend-dev / @qc-specialist / @qa-engineer 角色职责边界
mstar-coding-behavior实现角色消费 DESIGN.md 前的通用编码约束
mstar-design-system一个实现需要收敛到既有 baseline(原型 / 设计导出 / 既有产品标记 / 书面契约)时:baseline 定位与机器化、三层 parity、defect/decision/phantom 分诊、coverage 证据

Scope (DESIGN.md lifecycle)

TopicSee
Normative spec: section definitions, token naming, light/dark rules, YAML frontmatter structurereferences/design-md-spec.md
Three-level completeness checklist (MVP / Standard / Production)references/completeness-checklist.md
Vercel Geist DESIGN.md as annotated referencereferences/vercel-example.md
Full template with YAML frontmatter and Level 2/3 placeholderstemplates/DESIGN.md.template
Dark theme template (same token names, different values)templates/DESIGN.dark.md.template

YAML frontmatter is the SSOT for token values. Colors, typography, spacing, rounded, and components live in the frontmatter as structured, machine-readable data. The Markdown body is supplementary documentation (rules, intent, usage guidance). When reading DESIGN.md, always parse the YAML frontmatter first; when writing, keep frontmatter and body in sync.

Out of scope: rendered UI preview tooling outside the repo; frontend implementation that consumes DESIGN.md tokens (use @frontend-dev / @fullstack-dev); QC review verdict rules (→ mstar-review-qc).

Location

  • Primary: project root DESIGN.md (human + agent visible, aligns with AGENTS.md)
  • Dark theme: project root DESIGN.dark.md (same token names, different values)
  • DESIGN.md is a project-level design contract, not a harness internal artifact. It lives beside README.md and AGENTS.md.

Prototype vs document

When the repository also ships a prototype, this document is a summary of that prototype, not a second source:

  • Authority. The prototype is the design; DESIGN.md states in prose what it establishes. On conflict the prototype wins — unless the prototype is silent, in which case the decision and its reason belong in DESIGN.md. An unrecorded silence becomes a guess in the next implementation.
  • One-directional sync. baseline → fixture → document. Change the prototype, rerun the generator that derives the machine-readable fixture, then bring DESIGN.md in line. Hand-editing the fixture, or iterating in DESIGN.md and back-filling the prototype, is the reverse direction and a defect regardless of which side turns out to be right.
  • The fixture must be a tracked file derived from a tracked source. A deriver or comparator that reads a gitignored working path at runtime has no baseline: on a clean checkout it fails, and in the degraded mode it warns and exits 0, which reads as a pass forever.
  • A lagging document fails nothing. A fixture that lags the prototype fails its check; nothing fails when DESIGN.md lags both. That direction is a review obligation: the reviewer compares the document against the fixture, because no gate will.

Converging an implementation onto that baseline — baseline kinds, parity layers, defect / decision / phantom triage — → skill mstar-design-system.

Role lifecycle

Creator: @architect (primary) + @product-manager (requirements)

@architect owns DESIGN.md content — token selection, naming, completeness level decisions. @product-manager provides design intent: brand identity, target audience, must-have UI patterns, accessibility requirements.

Orchestrator: @project-manager

In Prepare phase, PM decides whether the project needs a DESIGN.md. If yes, dispatches to @architect with product requirements from @product-manager. PM checks DESIGN.md exists and meets the assigned completeness level before plan(locked).

Consumers

  • @frontend-dev / @fullstack-dev — read DESIGN.md before implementing styled components; map tokens to CSS/theme variables
  • @qc-specialist — verify UI implementation aligns with DESIGN.md tokens
  • @qa-engineer — verify visual output matches design spec when QA gate: mandatory

Phase gate integration

DESIGN.md is a Prepare-stage artifact (like spec). It must be created and reviewed before plan(locked) for any plan that includes UI work.

  1. PM includes "DESIGN.md creation/audit" in Prepare tracking checklist when the plan involves UI
  2. @architect creates or updates DESIGN.md; @product-manager reviews design intent alignment
  3. PM gates on: DESIGN.md exists, meets completeness level declared in plan, @product-manager signed off

For hotfix or plans with no UI changes, DESIGN.md check may be skipped.

Completeness levels(Evidence)

DESIGN.md supports three levels, each with built-in upgrade path:

  1. Level 1 — MVP (minimal, prevents guesswork): palette, base typography, spacing scale
  2. Level 2 — Standard (consistent components): full token scales, breakpoints, component tokens (Button, Input)
  3. Level 3 — Production (complete design system): dual theme, elevation, motion, shapes, component library, voice

The template includes all levels; Level 2 and 3 sections are commented out with <!-- LEVEL2_PLACEHOLDER: ... --> markers that explain when to activate them. The audit workflow detects these placeholders and can recommend upgrade.

Full checklist → references/completeness-checklist.md.

Workflows

Workflow 1: Create DESIGN.md (Prepare phase)

  1. Read references/design-md-spec.md for section definitions and YAML frontmatter structure
  2. Copy templates/DESIGN.md.template to {PROJECT_ROOT}/DESIGN.md — the template includes the full YAML frontmatter skeleton
  3. Interview @product-manager for brand colors, typography preferences, must-have patterns
  4. Fill Level 1 frontmatter tokens (uncomment and replace "[placeholder]" values with concrete hex/px values)
  5. If plan requires Level 2+ out of the gate, uncomment and fill those sections too
  6. Update the body prose to match the frontmatter values (target audience, aesthetic principles, rhythm rules)
  7. Run the completeness audit workflow below to confirm level
  8. Report to PM: path created, level achieved, what's needed for next level

Workflow 2: Audit DESIGN.md completeness

  1. Read DESIGN.md and DESIGN.dark.md (if exists) — parse the YAML frontmatter for structured token values
  2. Load references/completeness-checklist.md
  3. Check each checklist item; note gaps in both frontmatter (missing/uncommented keys, placeholder values) and body (missing rules/documentation)
  4. Report:
    • Current completeness level
    • Gaps preventing next level (frontmatter gaps vs. body gaps, tagged separately)
    • Presence of upgrade placeholders (LEVEL2_PLACEHOLDER, LEVEL3_PLACEHOLDER in both frontmatter comments and body HTML comments)
    • Recommendation: whether to upgrade now or defer
  5. Update DESIGN.md level tag (e.g., <!-- COMPLETENESS_LEVEL: 1 — last audited YYYY-MM-DD -->) if changed

Workflow 3: Add dark theme

  1. Read existing DESIGN.md to extract token names
  2. Copy templates/DESIGN.dark.md.template to {PROJECT_ROOT}/DESIGN.dark.md
  3. For each token in DESIGN.md, define the dark-theme equivalent value
  4. Preserve same token names; only values change (see references/design-md-spec.md § Light/Dark rules)
  5. Audit with Workflow 2 to confirm Level 3 completeness

Workflow 4: Consume DESIGN.md (implementation roles)

Before writing styled UI code:

  1. Read DESIGN.md (and DESIGN.dark.md if exists)
  2. Parse the YAML frontmatter for token values — this is the SSOT for colors, typography, spacing, rounded, and components
  3. Resolve component {colors.X}, {typography.X}, {rounded.X} references by tracing back to the corresponding frontmatter keys
  4. Extract tokens into implementation layer (CSS custom properties, Tailwind config, theme object, etc.)
  5. Follow DESIGN.md body Voice & Content rules for copy text
  6. If DESIGN.md is missing, has no frontmatter, or is incomplete, report to PM — do not guess tokens

Light/Dark dual-theme rules

Dual theme uses same token names, different values across two files:

DESIGN.md         DESIGN.dark.md
-----------       --------------
gray-100: #fff    gray-100: #111
gray-1000: #000   gray-1000: #eee
  • Token names are the SSOT interface — consumers reference tokens by name, not raw values
  • references/design-md-spec.md § Light/Dark rules defines the contract

Engine check (when available): run mstar design-md validate <dir> (or import { validateDesignTokenFrontmatter, assertLightDarkParity, completenessLevel } from "@mstar-harness/engine" in a host hook) to validate the token frontmatter schema above, the light/dark parity of DESIGN.dark.md (same token names, different values), and the achieved completeness level. On fail -> do not proceed; fix and re-run. Skill text below remains authoritative when the runtime is absent.

References

  • references/design-md-spec.md — normative spec: section definitions, token naming conventions, light/dark contract
  • references/completeness-checklist.md — three-level audit checklist with detailed criteria per level
  • references/vercel-example.md — Vercel Geist DESIGN.md as annotated reference (read when creating from scratch or needing design inspiration)

Templates (this skill):

  • templates/DESIGN.md.template — full template including all Level 1-3 sections with placeholder comments
  • templates/DESIGN.dark.md.template — dark theme template with same token names, different values

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Mstar Design Md AI skill do?

DESIGN.md 设计系统规范:创建/审计/维护 design tokens、三级完整性清单、light/dark 双主题(DESIGN.dark.md)。用户提到 DESIGN.md / design tokens / design system,或实现、审查 UI 时读取。

Why use Mstar Design Md on TypingMind?

Because you install it once and use it with any model. Mstar Design Md is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Mstar Design Md in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/btspoony/mstar-harness/tree/main/skills/mstar-design-md. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Mstar Design Md?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Mstar Design Md?

As many as you like. As long as a model supports skills, you can use Mstar Design Md with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Mstar Design Md AI skill free?

Yes. It is published on GitHub by btspoony under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇