Contributing to Claude Code Superpowers
Contributions are welcome and encouraged. This document explains how to add new skills, improve existing ones, and submit pattern files.
Loading actions...
Skill content
Main instructions and any bundled files for this skill.
Contributing to Claude Code Superpowers
Contributions are welcome and encouraged. This document explains how to add new skills, improve existing ones, and submit pattern files.
Types of Contributions
New domain skills
Skills for domains not yet covered: DevOps/Kubernetes, security engineering, data engineering, mobile (iOS/Android), game development, etc.
New pattern files
Code examples and configuration templates for existing skills. If a skill's SKILL.md references "load patterns/X.md" but the file does not exist, that is an open contribution opportunity.
Improved entry points
Better assessment questions that route more accurately to the right section.
Community custom skills
Skills your team built for your domain — share them so others benefit.
Bug fixes
Incorrect patterns, broken code examples, outdated configurations.
Skill Design Principles
Before writing, read one existing SKILL.md to understand the format. Key principles:
Every skill must have an entry point. Never dump all patterns immediately. Ask what stage or type of work before applying anything.
Distinguish rigid from flexible. Rigid skills (TDD, debugging, verification) are followed exactly every time. Flexible skills (domain expertise) adapt to context. Know which yours is.
Red flags are mandatory. The "Never" and "Always" section prevents the most common mistakes. If you cannot list at least 3 red flags, the skill is not specific enough.
Final checklist = completion gate. Every checklist item must be verifiable. "Code is good" is not a checklist item. "All tests pass and P99 latency is under budget" is.
Extract heavy code to patterns/. If a section contains more than ~20 lines of code or configuration, move it to patterns/<section-name>.md and reference it: "Load patterns: patterns/<section-name>.md".
File Structure
skills/
your-skill-name/
SKILL.md ← required, lean (target <900 tokens)
patterns/ ← optional, created if SKILL.md references patterns
section-one.md
section-two.md
SKILL.md Template
---
name: your-skill-name
description: Specific one-line description of WHEN this skill applies. Be concrete.
type: process | domain | implementation
---
# Your Skill Title
## Overview
**Core principle:** One sentence capturing the essential insight of this domain.
This skill provides **[what it provides]** for **[who uses it]**.
**Announce at start:** "I'm using the [name] skill for [purpose]."
---
## Entry Point — First 5 Minutes
When invoked, determine [what you need to know] before applying any patterns:
\`\`\`
ASSESSMENT:
"[Question that determines which section applies]"
A) Option A
B) Option B
C) Option C
→ Go to the relevant section.
\`\`\`
**Option → Section mapping:**
- A → [Section name]
- B → [Section name]
- C → [Section name]
**After identifying [dimension], ask:**
"[The one question whose answer shapes all trade-offs]"
---
## Section A: [Name]
[Content, or: Load patterns: `patterns/section-a.md`]
---
## Red Flags
**Never:**
- [Specific thing to never do]
- [Another specific thing]
**Always:**
- [Specific thing to always do]
- [Another specific thing]
---
## Integration with Superpowers
**Used with:**
| Skill | Integration |
|-------|-------------|
| `test-driven-development` | [How] |
| `verification-before-completion` | [How] |
---
## Final Checklist
Before claiming [domain] work complete:
- [ ] [Verifiable criterion]
- [ ] [Verifiable criterion]
- [ ] [Verifiable criterion]
Pattern File Template
# [Domain] — [Section Name] Patterns
## [Pattern Group 1]
\`\`\`[language]
[code example]
\`\`\`
## [Pattern Group 2]
\`\`\`[language]
[code example]
\`\`\`
Submission Process
- Fork the repository
- Create a branch:
git checkout -b skill/your-skill-name - Add your skill directory under
skills/ - Test it:
- Install with
bash install.sh - Open Claude Code in a project
- Invoke your skill:
/your-skill-name - Verify the entry point routes correctly
- Verify the red flags section is actionable
- Install with
- Open a PR with:
- Title:
skill: add [skill-name] — [one-line description] - Body: What domain it covers, what failure modes it prevents, example invocation
- Title:
Skill Quality Bar
A PR will be merged if the skill:
- Has a specific
descriptionthat enables accurate relevance matching - Has an entry point that asks before dumping patterns
- Has a Red Flags section with ≥ 3 concrete never/always items
- Has a Final Checklist with ≥ 3 verifiable items
- Keeps SKILL.md under ~900 tokens (move code to patterns/ if larger)
- Includes a code example (inline or in patterns/) that compiles/runs
Questions
Open an issue or start a discussion. Community questions help improve the documentation.
Related Skills
Frontend Typescript Linting.mdc
TypeScript and ESLint rules that MUST be followed when creating, modifying, or reviewing any file under apps/frontend/, including .ts, .tsx, .js, and .jsx files. Also apply when discussing frontend li...
2. Apply Deepthink Protocol (reason about dependencies
risks