Emil Kowalski Skills - Setup Guide¶

Source: skills.sh (694.4K combined installs) GitHub: emilkowalski/skills Category: Design Engineering First Seen: August 13, 2026 sweep Quality Tier: 🟢 Production

Emil Kowalski is the creator of Sonner (the toast library) and vaul, and one of the most-copied voices in agent UI quality. His skill pack is the highest-signal design-quality layer on skills.sh: it teaches agents to recognize bad AI-generated UI, audit motion instead of sprinkling it, and pick the right library before writing a component. emil-design-eng is the flagship at 204.4K installs.

Installation¶

npx skills add emilkowalski/skills

Installs all 10 skills. Hermes: place the SKILL.md files under your agent's skills directory, or run inside any skills.sh-compatible agent (Claude Code, Codex, Cursor, Hermes, OpenClaw, Nous Research).

Core Skills¶

Skill	Installs	Use For
emil-design-eng	204.4K	Fluid, tactile interface design - the core design-engineering doctrine
review-animations	97.6K	Adversarial review of animation/motion code against craft bars
animation-vocabulary	85.1K	Reverse-lookup glossary: vague descriptions → named motion primitives
apple-design	77.0K	Apple's HIG-informed approach to fluid, physical interfaces
improve-animations	70.1K	Senior-engineer pass over a codebase's animation and motion code
find-animation-opportunities	58.7K	Scan a UI for places that should animate but don't
pick-ui-library	44.8K	Choose the right frontend library for the task before building
prototype	34.0K	Build multiple genuinely different UI versions of a piece
animate	16.8K	Build an animation from scratch, making the decisions
ask-sonner	5.9K	Sonner-specific component guidance

Prerequisites¶

A frontend project (React-based for most skills; ask-sonner requires Sonner)
Access to motion libraries (Framer Motion / Motion One) for animation skills

CorpusIQ Use Cases¶

UI quality gate for agent-built dashboards - run review-animations + improve-animations before shipping any agent-generated frontend, the same way anti-ui-slop gates visual design
Docs site polish - find-animation-opportunities on docs.corpusiq.io surfaces motion gaps in landing and onboarding flows
Design-vocabulary alignment - animation-vocabulary gives team prompts shared language when briefing video/graphics work (pairs with the HyperFrames pipeline)

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects

Limitations / Verification¶

Quality-tier judgement is subjective - the skills encode opinions, not measurements; treat verdicts as review feedback, not tests
Verify installation with ls ~/.claude/skills/ | grep -i emil or your agent's equivalent skills directory listing