Skill Engineering — Design skill for Claude Code
Stop collecting skills.
How to install Skill Engineering
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open cobusgreyling/skill-engineering and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Skill Engineering does
Stop collecting skills. Design the load. Get a Skill Ready score.
Alternatives in Design
- Figma Implement Design — Translate Figma designs into production-ready code with pixel-perfect accuracy 14.6k ★
- Interface Design — Design engineering for Claude Code 4.3k ★
- Pack — Bundle every designlang output (DTCG tokens, Tailwind, shadcn, Figma vars, motion, anatomy, Storybook, prompts 3.6k ★
README
Skill Engineering
**Stop collecting skills. Design the load. Get a score.**
The description of every skill is loaded on every turn, before the task starts. The body is loaded only when a skill is selected. A long description is not documentation. It is a tax.
The score is that tax. 100 means the catalog is inside the warn line. The level is selection: a trigger, a distinct name, a description inside the spec limit of 1024 characters. A folder can score 100 and still be L1.
npx @cobusgreyling/skill doctor .
Or from a checkout:
git clone https://github.com/cobusgreyling/skill-engineering.git
cd skill-engineering
node bin/skill.mjs doctor examples/messy
Grade a project you already have:
npx @cobusgreyling/skill doctor /path/to/your/repo
npx @cobusgreyling/skill init /path/to/your/repo
`doctor` reads `SKILL.md` files under `.grok/skills`, `.claude/skills`, `.cursor/skills`, `.agents/skills`, `.codex/skills`, and `skills/`. A directory of skill folders is graded directly. `--root` adds another one. `--user` adds the same folders in your home directory, plus `~/.grok/bundled/skills` and marketplace `skills` folders.
On a pull request:
- uses: cobusgreyling/skill-engineering@main
with:
path: .
min: L2
baseline: skill-baseline.json
Exit code is `0` at L2 or better. `--min L3` makes warnings fail CI. `--baseline` is a previous `doctor --json` report. The action comments the catalog delta.
This repository scores itself:
node bin/skill.mjs doctor .
Skill Ready L3 catalog 36/4000
score 100 · 1 skill
A folder that fails
`examples/messy` has an essay where a trigger should be, a second skill with the same name, and a description that never says when to run. The catalog is still sm
Related Skills
Growth Experiment
Design a growth experiment with hypothesis, ICE score, metrics, and success criteria. Ready to execute and mea
Claude Code Reverse Engineering
Reverse-engineered design documents for Anthropic Claude Code CLI - implementation-ready specifications
Brand Context
Load and validate brand assets before any design work — logo, colors, typography, voice
Visual Prompt Coach
Claude Code skill — design visuals for technical course materials using Dan Roam, Mayer, C4, cognitive load th
Af Ship From Tech Design
If $ARGUMENTS is empty, stop and ask:
Factory Init
Interactively design and scaffold an autonomous agent loop in this project (goal, gate, state, stop conditions
Related Agents
Lead Scorer
Gives verified leads a transparent 0–100 score, explains why each is a good lead and what to offer first, merg
Fact Check
Verify load-bearing empirical claims before they are taught. Dispatched by /teach during its plan phase for to
Engineering Email Intelligence Engineer
Expert in extracting structured, reasoning-ready data from raw email threads for AI agents and automation syst