claude-skills
miovrag/claude-skills
Claude Code skills and commands
Instalação
npx convez-ai miovrag/claude-skills1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
Agent Skills
Single source of truth for all agent skills (Claude Code & friends). Lives at ~/.agents/skills/;
~/.claude/skills/ contains only symlinks into this directory. Invoke any skill as /<folder-name>.
Provenance
Own — CustomGPT.ai design work
| Skill | What it does |
|---|---|
customgpt-design | CustomGPT.ai design system, strict token mode (system root: ~/customgpt-design-system/) |
pxo | Full 13-step Product Experience OS workflow analysis |
premortem | Structured premortem risk analysis (8 steps) |
Own — UX pattern library
Distilled from Vitaly Friedman's "Fixing Frustrating Design Patterns" workshop, Baymard, NN/g, and related sources.
| Skill | Area |
|---|---|
filtering-ux | Filter UIs — interaction, taxonomy, layout |
sorting-ux | Sorting — filter-vs-sort, faceted sorting, labels |
data-table-ux | Tables/lists/grids — layout, interaction, a11y |
navigation-ux | Nav, breadcrumbs, menus, IA |
web-forms-ux | Forms, validation, multi-step, a11y |
dashboard-dataviz | Dashboards, chart selection, KPI cards |
search-and-content-ux | Search, autocomplete, scanability |
voice-ux | Voice interfaces, conversation design |
ai-design-patterns | Human-AI interaction, trust, transparency |
prompt-engineering | Prompt design and refinement for LLMs |
The list-view trio (filtering-ux, sorting-ux, data-table-ux) is designed to load together
whenever a list/table/grid is in scope.
Own — analysis frameworks
| Skill | What it does |
|---|---|
bourdieu-distinction | Bourdieu's Distinction framework → 8-section sociological analysis |
Third-party (vendored)
| Set | Source | Skills |
|---|---|---|
| Layers of Product Design | jamiemill/layers-skills | layers-intro, layers-orient, layers-domain, layers-observed-behaviour, layers-user-needs, layers-product-strategy, layers-conceptual-model, layers-interaction-flow, layers-surface |
| Accessibility (Mike Gifford / CivicActions) | mgifford/accessibility-skills | 27 granular a11y skills: accessibility-general, color-contrast, keyboard, forms, tables, tooltips, navigation, aria-live-regions, image-alt-text, svg, charts-graphs, touch-pointer, light-dark-mode, plain-language, content-design, manual-testing, axe-rules, bug-reporting, ci-cd, audio-video, maps, mermaid, print, progressive-enhancement, user-personalization, anchor-links, opquast-digital-quality |
| Remotion | remotion-dev via npx skills | remotion-best-practices |
| skills.sh | vercel-labs/skills | find-skills, project-kickstart |
| Designer skills | Owl-Listener/designer-skills | ~93 skills across design-research, design-systems, interaction-design, design-ops, designer-toolkit (critiques, UX laws, tokens, research methods, specs) |
| PM skills | phuryn/pm-skills | 68 skills across product-discovery, product-strategy, execution, data-analytics, market-research, go-to-market, marketing-growth, ai-shipping, toolkit |
| Animate | delphi-ai/animate-skill | animate — React/Next.js animation patterns (Emil Kowalski's course) |
| Journey Mapping | cuellarfr/design-skills | journey-mapping — alignment diagrams (CJM, service blueprints, experience/empathy maps, JJG Visual Vocabulary) with references/templates/examples; single skill vendored from the set to avoid name collisions |
Name collisions: interview-script and summarize-interview exist in both designer-skills and
pm-skills — the pm-skills versions (installed later, Mom Test/JTBD-based) are the ones on disk.
Third-party sets are vendored copies — re-installable via npx skills add <source>, but tracked
here so local modifications survive.
Which analysis skill when
/pxo— full product-experience flow analysis of an existing screen/flow/premortem— pre-launch risk list ("it already failed — why?")/layers-orient— diagnose which design layer (research → strategy → model → flow → surface) is the bottleneck; then drill in with the matching/layers-*skill (load/layers-introfirst)