← Corpus / dididecks-ai / agent-skill
dididecks-ai/agent-skills/loops/maintain-design-md/templates/design-md-scaffold
★ One paragraph describing the visual identity. Mention the runtime source-of-truth path so future agents know where to read the actual values.
- Path
- agent-skills/loops/maintain-design-md/templates/design-md-scaffold.md
★ Project Name — Design System
The runtime source of truth is
★ src/path/to/theme-or-base-layout.css. This document is the human- and agent-readable contract that explains intent. Keep the two in sync when either changes.
Brand & Style
★ Describe the project’s emotional register. Playful or professional? Dense or spacious? What metaphors recur? What’s the build-in-public posture? What does an agent need to know to make a stylistic decision that feels like this project when no specific rule covers the case?
Colors
★ Prose explanation of the palette. Why these colors? How does each signal (primary = the action, accent = noteworthy)? How are borders tiered (default / strong / accent)? What’s the relationship between surface tones?
Typography
★ Why this family or families? What does each cue? What’s the reading-width convention? Are headlines mono or sans?
Layout & Spacing
★ Container model (fixed-max / fluid / grid)? Vertical rhythm rule? Grid-gap conventions? Centering convention (explicit margin-inline / margin shorthand)?
Elevation & Depth
★ Flat or shadow-heavy? What signals “higher in the stack” — tonal layers, borders, shadows, blur? Where is glow / shadow reserved?
Shapes
★ Rounded scale assignments — what kinds of elements get which radius? Border conventions? Icon stroke / fill conventions?
Components
★ One subsection per reusable component. Each should answer: what it looks like, what props/variants exist, what scope-specific overrides are documented, and any cross-component contracts (e.g. “the footer’s GitHub icon shares its 36×36 round shape with the share button”).
★ ComponentName
★ Prose for this component.
Do’s and Don’ts
★ Hard rules distilled from prior mistakes. Each bullet should pair the rule with its why (or a brief incident reference) so the next agent knows whether the rule still applies to a new edge case. Add a bullet whenever a visual mistake gets corrected.
- Do …
- Don’t …