Open Source AI Prompts: Publishing the Workflows That Actually Work
After writing about our AI-assisted development process, I realized something important: the prompts and workflows we’ve developed aren’t just useful for us—they could help other developers build better solutions faster.
So here’s the idea: publish our AI prompts as open source resources.
Why This Matters
Most AI prompts you find online are generic or theoretical. Ours are different because they’re:
- Battle-tested – Used in real development scenarios, not just experiments
- Context-aware – Include specific examples and troubleshooting
- Systems-oriented – Part of integrated workflows, not standalone tools
- Documented – With clear explanations of why they work
Think of it as systems building for AI assistance: we build the prompts so you can build better solutions in your domain.
What We’d Publish
WordPress Development Prompts
The workflows we use for content creation and management:
- Blog post creation with authentic voice
- Multi-brand content syndication
- WordPress block troubleshooting
- Mobile-first design principles
- SEO optimization workflows
MCP Server Configurations
The Model Context Protocol setups that actually work:
- WordPress MCP for automated publishing
- Discord MCP for community engagement
- Draw.io MCP for documentation
- Search MCPs for research workflows
Development Workflows
The processes we use for AI-assisted development:
- Error handling and debugging prompts
- Code review and optimization workflows
- Testing strategy prompts
- Documentation generation
The Structure
We’re thinking of organizing this as:
- GitHub Repository – “steelbak-prompts” with version control and community contributions
- Steelbak.com Section – Dedicated area with searchable prompt library
- Documentation – Each prompt includes context, examples, and troubleshooting
- Community – Others can submit their working prompts
Why Now?
The timing feels right because:
- AI tools are maturing – But good workflows are still hard to find
- Community needs this – Developers are struggling with ineffective prompts
- We have proven results – Our FarrierHub development shows these workflows work
- Open source momentum – Aligns with our systems building philosophy
What We Need From You
Before we build this out, we want to know:
- What prompts would be most valuable? – What workflows are you struggling with?
- How would you want to use them? – GitHub? Website? Documentation format?
- What’s missing? – What AI assistance resources do you wish existed?
- Would you contribute? – Do you have working prompts to share?
The Vision
Imagine a world where developers don’t have to figure out AI workflows from scratch. Where proven prompts for common tasks are freely available. Where the community shares what actually works, not just what sounds good.
That’s what we want to build. Not another generic prompt collection, but a curated library of workflows that actually work.
It’s systems building for AI assistance: we build the tools and processes so you can build better solutions in your domain.
Next Steps
If this resonates with you:
- Join the conversation – Drop into our Discord and share your thoughts
- Follow along – We’ll post updates as we develop this
- Share this post – Help us reach other developers who need this
- Start thinking – What prompts would you want to see?
The tools are open source. The approach is documented. The community is growing.
Let’s build the AI workflows that actually work.
Related Resources
- How We’re Using AI to Build FarrierHub – The post that started this idea
- What is Steelbak? – Our systems building philosophy
- The Hub – Our community and collaborative projects
- Model Context Protocol – The MCP framework we use
- GitHub – Our open source repositories
- Discord – Join the conversation