DESIGN.md is a structured markdown file that defines your project's brand spec — tokens (colors, typography, spacing, radius), voice, narrative, principles, personas, states, and motion. AI coding agents (Claude Code, Codex, OpenCode, Cursor) read it as authoritative context before generating any UI, so the output looks like your brand instead of generic 'modern minimalist' slop.

How does oh-my-design work?

Run `npx oh-my-design-cli install-skills` once in your project. It installs 17 skills (core flow + live-capture + asset + the v0.2 agent layer: orchestrator, kr-writer, locale-adapter, designer-review, final-qa, codex-image), 16 sub-agents (master orchestrator + 15 specialists), 4 hooks, and 221 reference DESIGN.md files. After restarting your AI coding agent, you just talk in natural language — skills route to the right sub-agents automatically.

Does it call any AI provider during install?

No. The install is a pure markdown copy — zero API calls, zero data leaves your machine. Your AI coding agent (Claude Code, Codex, OpenCode, or Cursor) is the inference layer; install just teaches it where to look.

Which AI coding agents are supported?

Claude Code, Codex, OpenCode, and Cursor — Cursor has a dedicated install channel (--agent cursor) that writes the .cursor/rules shim plus the shared reference catalog. The skill markdown files are agent-agnostic; hooks ship for Claude Code specifically.

What is the design harness?

The omd-harness skill runs a 10-phase design pipeline (Discovery → Research → IA → Wireframes → Design system → Components → Asset sourcing → Microcopy → Validation → Handoff). It dispatches sub-agents in parallel where independent, asks 3 mandatory user checkpoints, and emits a v0/Cursor-ready package. Trigger it by saying things like 'design the entire onboarding from scratch' or '/omd-harness '.

Yes — MIT licensed, open source, no signup, no API key, no paid tier. The references in the bundle belong to their respective companies and are reproduced for educational reference only.

oh-my-design v1.8.0 (2026-06-10)

The taste loop closes: your corrections become preferences, preferences become proposals, and one click folds them into DESIGN.md. Plus a taste dashboard and a scripted release bench.

Auto-fold gate. When a preference scope recurs past the threshold, session-end-foldin writes a structured proposal and your next session opens by asking — one selectable-option question ("이 취향, DESIGN.md에 반영할까요?"). Approve → omd:learn folds it; decline → snoozed until it recurs again. The old remember → wait → manually-run-learn dance is gone.
Ambient capture now persists — and sees more. post-edit-watch no longer just warns: off-system drift is recorded to .omd/preferences.md as confidence: inferred (24h dedup, never while editing DESIGN.md/.omd). Detection grew from hex colors to radius and motion-duration — including Tailwind classes, kebab-case CSS, camelCase JSX inline styles, and .html outputs (what omd:harness actually emits).
omd:taste — see what the system thinks you like. One command renders applied / pending (with recurrence counts and how far from the fold threshold) / snoozed / not-yet-known axes, with inline actions (fold now · forget · refine). Skill count: 16 → 17.
Reviewers feed the loop. omd:designer-review and omd:final-qa now end recurring findings with a single opt-in question — "이 패턴, 취향으로 기록할까요?" — recording with signal: review on consent. Never auto-recorded, never auto-folded.
Scripted 2-bench release routine. scripts/bench/score-init.mjs (first-60s) and the new scripts/bench/score-taste.mjs (capture → reflect → propose) score real headless runs deterministically; both at 10/10 on this release.
Fixes & hygiene: session-state-loader output envelope corrected (its session-start context was silently dropped); catalog resolution order normalized across all 4 documents with a CI drift guard; installer channel mapping unified into one table (cursor dedup behavior preserved, tested); omd:kr-writer presets now self-contained for npx installs; dangling dev-only references cleaned; docs/JSON-LD FAQ generated from one source; release workflow actions bumped for Node 24 runners.