skill
claude-skill-pack
Claude skills organized into domain layers — engineering, product, founder/GTM, design, and documents
Set up
Skills are installed by placing their folder where your agent loads skills.
The download comes from github.com/BTN101/claude-skill-pack itself, not from our server, so what you get is exactly what its author published.
Install instructions come from the README
README
synced Oct 23, 2026claude-skill-pack
A curated collection of Claude skills organized into domain layers — covering engineering, product thinking, founder/GTM work, design quality, and document creation.
Built for use with Claude's skill system on claude.ai.
What are Claude Skills?
Skills are structured instruction sets that tell Claude how to approach a specific task — what workflow to follow, what quality gates to apply, what patterns to enforce. They're loaded into context when a matching task comes up, giving Claude specialist-level discipline rather than generic help.
Each skill lives in a folder with a SKILL.md file. The frontmatter description field is what triggers the skill — Claude reads it and decides whether to apply the skill based on your request.
skill-name/
├── SKILL.md # Required: name, description, instructions
└── references/ # Optional: supporting docs loaded on demand
Skill Layers
This pack introduces layer skills — meta-skills that group related specialists together. Instead of needing to know which exact skill to reach for, you describe the task and the layer routes to the right one automatically.
| Layer | Domain | Skills Inside |
|---|---|---|
founder-layer | Business, investors, GTM, content | investor-materials, investor-outreach, market-research, brand-voice, copywriting, article-writing, launch-strategy, pricing-strategy, competitive-landscape |
engineering-layer | Code, architecture, APIs, security | api-design, backend-patterns, security-review, react-patterns, react-performance, nextjs-app-router-patterns, postgres-patterns, database-migrations, deployment-patterns, tdd-workflow, java-coding-standards |
security-layer | Security auditing, vulnerability hunting, vibe-coded app hardening | security-review, find-bugs, gha-security-review, code-review |
product-layer | Product strategy, discovery, specs | product-lens, lean-startup, 37signals-way, mom-test, jobs-to-be-done, prioritization-advisor, prd-development, hooked-ux, user-story |
design-layer | UI quality, motion, anti-slop | impeccable, design-taste-frontend, redesign-existing-projects, framer-motion-animator, animation-micro-interaction-pack |
document-layer | File creation and reading | docx, pptx, xlsx, pdf, file-reading, pdf-reading |
How It Works
You: "Cold email an investor about my fraud detection startup"
↓
Layer: founder-layer triggers on "investor" + "email"
↓
Skill: investor-outreach runs its structured workflow
↓
Output: A tight, personalized, proof-first email — not generic copy
The layer skill is the router. The individual skills are the knowledge. Both need to be installed.
Skills in This Pack
🟠 Layers (skills/layers/)
| Skill | What it does |
|---|---|
founder-layer | Routes business, investor, GTM, and content tasks |
engineering-layer | Routes code, architecture, security, and deployment tasks |
security-layer | Routes security audits, vulnerability hunts, and vibe-coded app hardening |
product-layer | Routes product strategy, discovery, and spec tasks |
design-layer | Routes UI quality, motion, and design audit tasks |
document-layer | Routes file creation and reading tasks |
⚙️ Engineering (skills/engineering/)
| Skill | What it does |
|---|---|
security-review | Secrets management, input validation, SQL injection, XSS, CSRF, auth, RLS |
find-bugs | Branch-diff scanner — maps attack surface, runs OWASP checklist per changed file |
gha-security-review | GitHub Actions audit — pwn-request, expression injection, supply chain, credential theft |
code-review | Holistic PR review — runtime errors, N+1s, perf, backwards compat, security |
api-design | REST naming, status codes, pagination, versioning, error format |
backend-patterns | Repository + service layer, caching, background jobs, middleware |
react-patterns | Composition, compound components, custom hooks, context |
react-performance | Memoization, code splitting, Server Components, waterfall fixes |
react-testing | Testing Library, Vitest/Jest, MSW mocking, accessibility assertions |
postgres-patterns | Query optimization, indexing, connection pooling, RLS |
database-migrations | Safe schema changes, rollbacks, zero-downtime migrations |
deployment-patterns | Vercel, Docker, CI/CD, health checks, rollback strategy |
tdd-workflow | Test-driven development, 80%+ coverage target |
java-coding-standards | Spring Boot / Quarkus naming, immutability, streams, exceptions |
🧠 Founder (skills/founder/)
| Skill | What it does |
|---|---|
investor-materials | Pitch decks, one-pagers, financial models, accelerator applications |
investor-outreach | Cold emails, warm intros, follow-ups — proof over adjectives |
market-research | TAM/SAM/SOM sizing, competitive analysis, investor diligence |
brand-voice | Reusable voice profile from real source material |
article-writing | Blog posts, articles, technical essays, newsletters |
🔭 Product (skills/product/)
| Skill | What it does |
|---|---|
product-lens | Founder diagnostic — who is this for, what's the pain, what's the MVP |
Installation
Install a .skill file on claude.ai
- Download the
.skillfile from the releases page - Go to claude.ai → Settings → Skills
- Click Install skill and select the
.skillfile - The skill is now active across your conversations
Install from source
Clone the repo and use any SKILL.md directly — the file format is plain markdown with YAML frontmatter.
git clone https://github.com/BTN101/claude-skill-pack
Skill File Format
Every skill follows this structure:
---
name: skill-name
description: When to trigger and what the skill does. This is the
primary mechanism — Claude matches your request against
all skill descriptions to decide what to load.
---
# Skill Name
Instructions for Claude go here in plain markdown.
## When to Activate
## Workflow
## Quality Gate
The description field is the most important part. It acts as the trigger — so it needs to be specific about both what the skill does and when it should fire.
Contributing
Skills are just markdown files. To add one:
- Create a folder under the relevant category in
skills/ - Add a
SKILL.mdwith proper YAML frontmatter - Optionally add a
references/folder for supporting docs - Submit a PR
Good skills have: a precise trigger description, a clear workflow, and a quality gate before output.
Why This Exists
Most AI assistance is generic. Skills make it structural — the same way a senior engineer follows a checklist before shipping an API endpoint, or a founder applies a framework before deciding what to build. This pack is an attempt to encode that discipline into reusable, shareable form.
License
MIT — use freely, modify as needed, attribution appreciated.
Built by @BTN101 · Centurion, South Africa