A Claude Code skill that generates LinkedIn and Facebook posts from any GitHub project — saved as a styled Word document you can open, copy, and paste directly.
Point it at any GitHub repo and it:
- Pulls your project data (README, metadata, languages, releases)
- Finds screenshots in your local project directory
- Generates detailed, emoji-rich posts for LinkedIn and Facebook
- Creates a styled
.docxfile in your project'ssocial-media-posts/folder - Embeds your project screenshots directly in the document
Open the Word file, copy the content for your platform, paste, attach images, post. That's it.
The generated Word document includes:
- Quick Guide — how to use the document
- LinkedIn Post — professional storytelling tone, with platform tips
- Facebook Post — casual and conversational, link in body
- Screenshots — your project images embedded and ready to attach
- Visual Tips — what GIF to record, what to screenshot, social preview suggestions
- Claude Code CLI installed
- GitHub CLI (
gh) authenticated - Python 3.11+ with
python-docx:
pip install python-docxUser-level (available in all projects):
# Clone to your Claude skills directory
git clone https://github.com/Ublaze/github-showcase.git ~/.claude/skills/github-showcaseProject-level (available only in one project):
git clone https://github.com/Ublaze/github-showcase.git .claude/skills/github-showcaseIn any Claude Code session:
/showcase Caffeinator
/showcase https://github.com/user/repo
The Word document will be saved to:
{your-project-dir}/social-media-posts/{ProjectName}_Showcase_{date}.docx
- The skill instructs Claude to gather your GitHub repo data via
gh api - Claude finds screenshots in your local project's
assets/,screenshots/, ordocs/images/folders - Claude generates platform-specific content (LinkedIn + Facebook) with emojis, hooks, and CTAs
- The
showcase_generator.pyscript creates a styled Word document with:- Boxed content sections (copy-paste ready)
- Platform-specific tips
- Embedded screenshots
- Visual content recommendations
Posts are generated with:
- Problem-first hooks — never "I'm excited to announce..."
- Generous emojis throughout
- Detailed content (1200-2000 characters) — not a teaser, a full showcase
- Feature bullet points with emojis
- Technical highlights kept accessible
- Personal reflection on what you learned
- CTA + hashtags at the end
| Requirement | Details |
|---|---|
| Claude Code | Any recent version |
| GitHub CLI | gh authenticated with your account |
| Python | 3.11+ |
| python-docx | pip install python-docx |
MIT