Thank you for your interest. We welcome contributions across:
- 🌏 Translations — especially English (Global Edition), then ES/FR/JA/KO/ZH
- 📊 Regional benchmarks — any country/market with verified data sources
- 🛒 Dropshipping skills — for the upcoming Dropshipping module (Q2 2026)
- 🚀 Personal branding skills — for the upcoming Personal Branding module (Q3 2026)
- 🐛 Bug fixes — typos, broken links, outdated information
- 💡 New skills — must fill a clear gap and follow the v2.0 frontmatter standard
# 1. Fork the repo on GitHub
# 2. Clone your fork
git clone https://github.com/YOUR-USERNAME/ai-business-skills.git
cd ai-business-skills
# 3. Create a branch
git checkout -b feature/your-skill-name
# 4. Use the template
cp skills/en/_TEMPLATE.md skills/en/your-skill-name.md
# Edit the new file following the v2.0 standard
# 5. Test in Claude Code or any AI tool
# Make sure the skill activates on its triggers and produces structured output
# 6. Commit
git add skills/en/your-skill-name.md
git commit -m "feat(skills): add your-skill-name (Global Edition)"
# 7. Push and open a Pull Request
git push origin feature/your-skill-nameEvery new skill must include:
- ✅ Frontmatter following v2.0 standard (see
skills/en/_TEMPLATE.md) - ✅ Clear
triggers:list (3–7 trigger phrases matching user intent) - ✅ Information-gathering protocol (max 4 questions)
- ✅ Structured output template (tables, not pure prose)
- ✅ Cross-references to related skills
- ✅ Quality checklist (built-in self-review)
- ✅ At least 1 example output (in
examples/or inline) - ✅ Under 500 lines — move long lookup tables, platform checklists, and extended templates into
skills/<lang>/<skill>/references/*.mdand link to them from a pointer line (progressive disclosure; see34-ai-marketing-osfor the pattern)
Run ./validate-skills.sh (or .\validate-skills.ps1 on Windows) before opening a PR. Target: Issues: 0.
Vietnamese skills write their body in unaccented Vietnamese (technical terms stay in English: ROAS, CPL, TOFU/MOFU/BOFU). This is a repo-wide convention that avoids encoding issues when skills are loaded into different AI platforms.
When translating Vietnam Edition → Global Edition:
- ✅ Frameworks (5 Whys, JTBD, RACI, persona templates, funnel stages)
- ✅ Methodology (question protocol, output structure, quality checklist)
- ✅ Cross-reference logic
- 🔄 Currency — VND → USD with regional context (not just exchange rate; rebase to local purchasing power)
- 🔄 Benchmarks — Pull from
references/benchmarks-global.mdinstead of literal VN numbers - 🔄 Channels — Zalo OA → Email/SMS; Shopee VN → Shopify/Amazon; TikTok Shop VN → TikTok Shop US/UK
- 🔄 Examples — Spa VN → DTC skincare US; F&B VN → DTC food brand US/UK
- 🔄 Compliance — VN advertising law → FTC, GDPR, CCPA references
- 🔄 Cultural context — VN buying behavior (COD-heavy, Zalo trust) → US/EU (credit card, review-driven)
If you add regional benchmarks (e.g., for a country not yet covered):
- Cite at least 2 independent sources with publication dates
- Use medians, not averages (unless explicitly noted otherwise)
- Specify date ranges (e.g., "Q4 2025 data")
- Format consistently with existing
references/benchmarks-*.mdfiles - Include both "poor/average/good/excellent" bands where applicable
- Update your branch with latest
master - Ensure your PR description explains: what, why, testing done
- Tag with relevant labels (
translation,new-skill,bug, etc.) - We review within 1 week
- Merged PRs get credit in
CONTRIBUTORS.mdand changelog
Be kind. Be specific. Be useful. Disagree on technical merit, never on identity.
Open a Discussion or reach out to @minhnv0807.