Where Does the Knowledge in the Emil Kowalski Skills Repository Come From?

The knowledge encoded in Emil Kowalski's skills repository comes from three sources: the author's personal design‑engineering philosophy and published works, curated external references like Apple's Human Interface Guidelines and easing curve libraries, and internal audit templates that codify professional standards.

The emilkowalski/skills repository is a curated knowledge base that ships as a CLI tool for design‑engineering workflows. Rather than generating advice algorithmically, it packages battle-tested principles into markdown "SKILL.md" files that the skills CLI reads at runtime. This article breaks down exactly where that knowledge originates and how it's structured in the codebase.

The Three Pillars of Knowledge

Understanding where the knowledge encoded in these skills comes from requires examining three distinct source categories woven throughout the repository.

Author's Personal Design Philosophy

The foundation of the skills repository is Emil Kowalski's own design‑engineering philosophy, drawn from his published essays and talks.

In skills/emil-design-eng/SKILL.md, core concepts appear verbatim from sources like the "You Don't Need Animations" essay and the "Agents with Taste" article. The README (README.md lines 15-19) quotes this same language, establishing restraint, polish, and invisible details as fundamental values.

This isn't abstract inspiration—the skill files extract specific, actionable guidance. The emil-design-eng skill serves as the primary interface for Kowalski's methodology, making his personal standards available programmatically.

Curated External References

The repository extensively references industry-standard resources rather than recreating knowledge:

• Easing curve libraries: Both easing.dev and easings.co are cited across multiple skills. The animate skill (SKILL.md line 112) and review-animations standards (STANDARDS.md line 37) explicitly direct developers to these tools instead of hand-rolling curves.

• Apple's design ecosystem: The apple-design skill (SKILL.md lines 49-55) distills WWDC talks and the Human Interface Guidelines into web-specific implementation guidance.

• Open-source component libraries: The pick-ui-library skill (SKILL.md lines 24-38) maintains a vetted list including base-ui, cmdk, Sonner, and motion.dev with direct links to each.

These references are not casual links—they're integrated into skill logic. When you run npx skills run pick-ui-library, the tool returns the curated list from this file.

Internal Audit Templates

The third knowledge source comprises domain-specific audit templates derived from Kowalski's internal quality checklists.

The improve-animations/AUDIT.md file (line 3) encodes concrete target values and evaluation criteria from real project audits. This transforms subjective quality standards into measurable checkpoints that the CLI can guide developers through.

How Knowledge Flows Through the Codebase

The repository structure reveals how these sources converge into usable skills:

File Path	Knowledge Function
README.md	Entry point; summarizes philosophy and installation
skills/emil-design-eng/SKILL.md	Core design‑engineering principles
skills/animation-vocabulary/SKILL.md	Controlled vocabulary for AI interpretation
skills/apple-design/SKILL.md	Platform-specific guidelines (Apple → web)
skills/pick-ui-library/SKILL.md	Curated library recommendations
skills/improve-animations/AUDIT.md	Quality audit checklist with target values

Each .md file in skills/ is both documentation and data source. The CLI tool parses these files at runtime, making the curated knowledge executable.

Practical Usage Examples

Installing and Running Skills

# Install the skill runner (from README)

npx skills@latest add emilkowalski/skills

# Generate an animation using curated easing knowledge

npx skills run animate "Add a 300 ms ease-out fade-in to the primary button"

The animate skill reads skills/animate/SKILL.md and references easing.dev when curve specifications are needed.

Selecting UI Libraries

npx skills run pick-ui-library "I need a toast component for a React app"

This executes against skills/pick-ui-library/SKILL.md, returning Sonner or alternatives from the vetted list with direct repository links.

Summary

The knowledge encoded in Emil Kowalski's skills repository originates from:

• Published works and talks by the author, quoted verbatim in skill definitions
• External standards including Apple HIG, easing libraries, and open-source UI kits
• Internal audit templates that codify professional quality benchmarks

All sources are packaged into markdown files that serve dual purpose: human-readable documentation and machine-parsable data for the skills CLI.

Frequently Asked Questions

Is the knowledge in the skills repository AI-generated?

No. According to the source code in emilkowalski/skills, all knowledge is human-curated. The skills extract Emil Kowalski's published essays, industry guidelines, and internal checklists into structured markdown files. AI systems consume this curated content—they don't create it.

Can I contribute my own skills to this repository?

The repository structure suggests skills are author-curated rather than community-driven. Each SKILL.md file contains specific attribution to Kowalski's philosophy or vetted external sources. There's no contribution template indicating open submissions.

How does the CLI tool use these markdown files?

The skills CLI parses the markdown files at runtime. Commands like npx skills run animate locate the corresponding skills/animate/SKILL.md, extract guidance, and execute workflows. The markdown serves as both documentation and executable configuration.

What makes this different from other design system tools?

Most design systems generate components programmatically. This repository encodes decision-making knowledge—when to animate, which easing to use, which library to pick. The improve-animations/AUDIT.md file exemplifies this: it doesn't generate code but guides quality evaluation using the author's professional standards.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}