Skip to content

ramybarsoum/claude-skills

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 

Repository files navigation

Claude Skills

110+ shareable skills for Claude Code. Built and maintained by the AllCare team.

We use Claude Code across product, engineering, and design. These skills capture workflows we've battle-tested internally, plus community skills we've found useful.

Install

One command. Installs all skills, configures hooks, done.

curl -fsSL https://raw.githubusercontent.com/ramybarsoum/claude-skills/main/install.sh | bash

Install a single skill:

curl -fsSL https://raw.githubusercontent.com/ramybarsoum/claude-skills/main/install.sh | bash -s -- feature-spec-interview

Other commands:

# Check for updates
curl -fsSL https://raw.githubusercontent.com/ramybarsoum/claude-skills/main/install.sh | bash -s -- --check

# Uninstall
curl -fsSL https://raw.githubusercontent.com/ramybarsoum/claude-skills/main/install.sh | bash -s -- --uninstall

Restart Claude Code after installing. Type /find-skills to browse everything.


Skills by Function

1. Product & Feature Specs

Everything a PM needs: spec writing, planning, strategy, research, and launch.

Skill Description
allcare-feature-spec-interview Full-stack spec interview framework. 14 prompts, 61 question groups, strategic gate, three-tier eval harness, tech design generation. Self-contained: all prompts, templates, and audit checklists in one file. GSD interactive execution model.
feature-spec-interview Interactive NLSpec interview. 5 modes, 40 question groups. Produces AI-agent-executable specs.
brainstorming Structured creative exploration before building anything.
writing-plans Create implementation plans from specs or requirements.
executing-plans Execute plans with review checkpoints.
doc-coauthoring Co-author documentation, proposals, technical specs, decision docs.
internal-comms Status reports, leadership updates, newsletters, FAQs, incident reports.
launch-strategy Product launch planning: phased launches, channel strategy, momentum.
copywriting Marketing copy for homepage, landing, pricing, feature, product pages.
copy-editing Systematic editing of marketing copy through multiple focused passes.
social-content Social media content for LinkedIn, Twitter/X, Instagram, TikTok.
email-sequence Drip campaigns, nurture sequences, onboarding emails, lifecycle programs.
marketing-psychology 70+ mental models for marketing: cognitive biases, persuasion, decision-making.
signup-flow-cro Optimize signup, registration, and trial activation flows.
onboarding-cro Post-signup onboarding, activation, first-run experience, time-to-value.
analytics-tracking GA4, GTM, conversion tracking, event tracking, UTM parameters.
ralph-wiggum Devil's advocate reviewer. Finds logical gaps and questionable assumptions with humor.
Plan Reviews (from gstack)
plan-ceo-review CEO/founder-mode plan review. 4 modes: Scope Expansion, Selective Expansion, Hold Scope, Scope Reduction. 11 review sections, 18 cognitive patterns.
plan-eng-review Eng manager-mode plan review. Architecture, data flow, edge cases, test coverage, performance. 15 cognitive patterns.
office-hours YC Office Hours. Startup mode: 6 forcing questions exposing demand reality. Builder mode: design thinking for side projects. Saves a design doc.

2. Coding

Code quality, review, testing, debugging, database, auth, and development workflows.

Skill Description
Quality & Clean Code
solid SOLID principles, TDD, clean code for production-grade software.
python-clean-code Full Clean Code catalog for Python: naming, functions, comments, DRY, boundaries.
boy-scout Boy Scout Rule: always leave code cleaner than you found it.
clean-names Descriptive names, appropriate length, no encodings.
clean-functions Max 3 arguments, single responsibility, no flag parameters.
clean-comments No metadata, no redundancy, no commented-out code.
clean-tests Fast tests, boundary coverage, one assert per test.
clean-general DRY, single responsibility, clear intent, no magic numbers.
code-simplifier Simplify code for clarity and maintainability. Preserves functionality.
Code Review
code-review Structured code review for PRs and implementation phases.
pr-review-toolkit Multi-agent PR review: code, silent failures, type design, test coverage.
frontend-code-review Review frontend files (.tsx, .ts, .js) against checklist.
requesting-code-review Verify requirements met before submitting for review.
receiving-code-review Process review feedback with rigor, not blind implementation.
verification-before-completion Run verification and confirm output before claiming done.
Testing & Debugging
test-driven-development TDD workflow. Use before writing implementation code.
systematic-debugging Scientific debugging: hypotheses, investigation, root cause.
webapp-testing Test local web apps with Playwright. Screenshots, browser logs.
Database & Backend
sql-optimization-patterns Query optimization, indexing, EXPLAIN analysis.
supabase-postgres-best-practices Postgres performance optimization from Supabase.
better-auth-best-practices Better Auth: comprehensive TypeScript authentication framework.
DevOps & Git
commit-commands Git commit, push, and PR creation workflows.
using-git-worktrees Isolated worktrees for feature work with safety verification.
finishing-a-development-branch Merge, PR, or cleanup when implementation is done.
safe-file-deletion Enforces explicit permission before any file deletion.
explain-code Explain code with visual diagrams and analogies.
PR & QA (from gstack)
review Pre-landing PR review. Two-pass analysis: critical (SQL safety, race conditions, LLM trust boundaries) + informational. Fix-First flow.
investigate Systematic debugging with root cause investigation. Iron Law: no fixes without root cause. 4 phases, 6 pattern signatures, 3-strike rule.
qa Full QA: test, fix with atomic commits, re-verify. Health score rubric across 8 categories. Three tiers: Quick, Standard, Exhaustive.
qa-only Report-only QA. Same methodology as qa but never fixes anything. Structured report with health score and screenshots.

3. UI/UX & Design

Frontend interfaces, design systems, browser automation, diagrams, and visual art.

Skill Description
Frontend Design
frontend-design Production-grade UI with high design quality. Avoids generic AI aesthetics.
frontend-ui-ux-engineer-skill Designer-turned-developer. Stunning UI/UX without mockups.
ui-ux-pro-max 50 styles, 21 palettes, 50 font pairings, 20 charts, 9 stacks.
ui-ux-designer Interface design, interaction design, user experience, design systems.
web-design-guidelines Review UI for Web Interface Guidelines compliance and accessibility.
interface-craft Polished animated interfaces in React. Storyboard DSL, DialKit, Critique.
vercel-react-best-practices React/Next.js performance optimization from Vercel Engineering.
Components & Theming
web-artifacts-builder Multi-component HTML artifacts with React, Tailwind, shadcn/ui.
canvas-design Visual art in .png and .pdf. Posters, art, static designs.
theme-factory 10 pre-set themes for slides, docs, reports, landing pages.
brand-guidelines Anthropic's official brand colors and typography.
motion Motion Vue animations: gesture, scroll-linked, layout transitions.
remotion-best-practices Video creation in React with Remotion.
Diagrams
mermaid-diagrams Software diagrams: class, sequence, flowchart, ERD, C4, state, gantt.
algorithmic-art Algorithmic art with p5.js: flow fields, particle systems.
Browser Automation
agent-browser Browser automation for testing, form filling, screenshots, extraction.
dev-browser Persistent browser state. Navigate, fill, screenshot, scrape.
Design Reviews (from gstack)
plan-design-review Designer's eye plan review. Rates each dimension 0-10, explains what would make it a 10, then fixes to get there. 9 design principles, 12 cognitive patterns.
gstack-design-review Visual QA audit. 80-item checklist across 10 categories. Finds spacing issues, hierarchy problems, AI slop patterns. Fixes with atomic commits + before/after screenshots.
design-consultation Full design consultation. Researches landscape, proposes complete design system, generates font+color preview pages. Creates DESIGN.md.

4. AI & Agents

Build AI features, MCP servers, and multi-agent systems.

Skill Description
agent-sdk-dev Claude Agent SDK application development.
mcp-builder Build MCP servers. Python (FastMCP) and Node/TypeScript.
context7 Retrieve up-to-date docs for any library via Context7 API.
feature-dev Feature development workflow with agents.
orchestrator Multi-agent orchestration.
dispatching-parallel-agents Run 2+ independent tasks in parallel.
subagent-driven-development Execute plans with independent subagent tasks.
get-shit-done GSD meta-prompting: plan, execute, verify with context-rot prevention.
para-memory-files PARA-method file-based memory: knowledge graph, daily notes, tacit knowledge, memory decay.
agent-heartbeat-pattern Structured execution loop for agents that wake periodically. Checkout, incremental context, blocked-task dedup.
agent-token-optimization Reduce token consumption in agent systems. Four root causes and systematic fixes from 11K+ production runs.

5. Mobile & Native

iOS, macOS, and cross-platform mobile development.

Skill Description
Swift & iOS
swift-expert iOS/macOS/watchOS/tvOS. SwiftUI, async/await, actors, protocols.
ios-swift-development MVVM, SwiftUI, URLSession, Combine, Core Data.
swift-style Swift code style: formatting, naming, organization.
swift-testing-expert Swift Testing: #expect/#require, traits, parameterized tests, async.
swift-protocol-di-testing Protocol-based DI for testable Swift. Mock file system, network, APIs.
swift-mcp-server-generator Generate MCP server projects in Swift.
SwiftUI
swiftui-expert-skill State management, composition, performance, iOS 26+ Liquid Glass.
swiftui-animation Animations, transitions, matched geometry, Metal shaders.
swiftui-liquid-glass iOS 26+ Liquid Glass API.
swiftui-performance-audit Diagnose slow rendering, janky scrolling, excessive updates.
swiftui-pro Comprehensive SwiftUI review: modern APIs, maintainability, perf.
swiftui-ui-patterns Example-driven guidance for views, components, TabView.
rivetkit-client-swiftui RivetKit SwiftUI client: Rivet Actors, @Actor, view modifiers.
Cross-Platform
building-native-ui Expo Router: styling, components, navigation, animations, native tabs.

6. Documents & Files

Create and manipulate office documents programmatically.

Skill Description
pdf Extract text/tables, create, merge/split, handle forms.
docx Word docs: tracked changes, comments, formatting.
pptx Presentations: layouts, speaker notes, content.
xlsx Spreadsheets: formulas, formatting, data analysis, visualization.
slack-gif-creator Animated GIFs optimized for Slack.

7. Workflow & Meta

Skills that modify how Claude Code itself operates.

Skill Description
using-superpowers Skill discovery and invocation framework.
find-skills Discover and install skills. "How do I do X?"
skill-creator Create new skills or update existing ones.
writing-skills Create, edit, and verify skills before deployment.
hookify Create hooks to prevent unwanted behaviors.
ask-user-question Ask users questions via UI for clarification.
learning-output-style Learning mode with educational explanations.
explanatory-output-style Educational insights about the codebase as you work.
claude-opus-4-5-migration Migration utilities for Claude Opus 4.5.
security-guidance Security hooks and guidance.
file-permission File permission enforcement.
Ship & Release (from gstack)
ship Ship workflow: merge base, run tests, review diff, bump VERSION, update CHANGELOG, commit, push, create PR. Fully automated.
document-release Post-ship docs update. Cross-references diff against all project docs. Updates README, ARCHITECTURE, CONTRIBUTING, CLAUDE.md.
retro Weekly engineering retrospective. Git analysis, work patterns, code quality metrics, team-aware praise and growth areas, streak tracking.

How Skills Work

Each skill is a folder under skills/ with a SKILL.md file and optional references/ directories. Claude Code reads the SKILL.md when invoked and loads reference files as needed.

Skills installed to ~/.claude/skills/ are globally available in every project.

Some skills are "plugins" with richer structure (commands/, agents/, hooks/). These provide slash commands, sub-agents, and automated hooks beyond the core skill.

Contributing

We welcome contributions from the AllCare team and the community.

Adding a new skill:

  1. Create a folder under skills/ with your skill name
  2. Add a SKILL.md with YAML frontmatter (name, description, category) and the skill instructions
  3. Optionally add a references/ directory for supporting files
  4. Update install.sh if the skill needs hooks or special setup
  5. Open a PR with a clear description of what the skill does and when to use it

Improving an existing skill:

  • Fix unclear instructions, add missing edge cases, improve question flows
  • Test by copying the updated SKILL.md to ~/.claude/skills/<skill-name>/ and running it

Guidelines:

  • Skills should be self-contained. No external dependencies beyond Claude Code.
  • Write instructions as if the reader has never seen the skill before.
  • Include concrete examples where possible.
  • Keep the scope focused. One skill, one job.

Team

Built by AllCare. Maintained by the product and engineering teams.

License

MIT

About

Shareable skills for Claude Code. Feature spec interviews, Dark Factory NLSpec methodology, and more.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors