This repository has been restructured as a comprehensive GitHub Pages site for better navigation and content extraction. The complete guide is now available at: https://danielmeppiel.github.io/awesome-ai-native
Programming has evolved. From Assembly to Python, each abstraction brought us closer to human thought. Now we've reached the final layer: prose itself becomes executable.
| Constraint | Principle | Induced Property | |
|---|---|---|---|
| P | Progressive Disclosure | Context arrives just-in-time, not just-in-case | Efficient context utilization |
| R | Reduced Scope | Match task size to context capacity | Manageable complexity |
| O | Orchestrated Composition | Simple things compose; complex things collapse | Flexibility, reusability |
| S | Safety Boundaries | Autonomy within guardrails | Reliability, security |
| E | Explicit Hierarchy | Specificity increases as scope narrows | Modularity, inheritance |
Use PROSE directly in Claude Code or GitHub Copilot CLI:
# Add the marketplace
/plugin marketplace add danielmeppiel/awesome-ai-native
# Install the skill
/plugin install prose-architect@proseOnce installed, the skill auto-activates when you:
- Ask to build an AI-native app from requirements
- Want to make a legacy project AI-native
- Need to design agent workflows or primitives
Access the full AI Native Development guide with improved navigation and structure
Three disciplines that implement PROSE: Prompt Engineering, Agent Primitives, Context Engineering
The architectural style definition—constraints, grounding principles, and derivation
Build your first Agent Primitives and see immediate results
Master advanced patterns and async delegation workflows
Scale AI Native Development across your organization
Checklists, progression guides, and documentation links
Your AI interactions are inconsistent and unreliable:
- Sometimes Copilot generates brilliant code, other times it's completely off-target
- You waste time re-prompting and re-explaining the same context repeatedly
- Different requests for similar tasks produce wildly different quality results
- Team members get different AI outputs for the same problems
- You can't predict or control what the AI will focus on
Sound familiar? You're experiencing the chaos of unstructured AI interaction.
Systematic approach to transform unreliable AI chats into consistent, professional workflows:
- Core Technique: Use structured Markdown to guide AI reasoning (like coding standards for prompts)
- Agent Primitives: Build reusable AI configurations that improve over time
- Context Engineering: Optimize AI memory and performance for complex projects
- Async Delegation: Scale through GitHub Coding Agents and multi-agent coordination
- Team Intelligence: Share successful AI patterns across your organization
Each PROSE constraint addresses a specific failure mode:
| Constraint | Failure Mode | Solution |
|---|---|---|
| Progressive Disclosure | Context overload dilutes attention | Load context just-in-time |
| Reduced Scope | Scope creep degrades quality | Right-size tasks to context capacity |
| Orchestrated Composition | Monolithic prompts collapse | Compose from small primitives |
| Scoped Boundaries | Unbounded autonomy is unsafe | Define tools, knowledge, approval |
| Explicit Hierarchy | Flat guidance pollutes context | Layer guidance global to local |
Traditional approach: "Tell the AI what to do"
PROSE approach: "Engineer the context and structure for optimal cognitive performance"
Ready to transform your AI development workflow? Visit the complete guide to choose your learning path and start building more reliable, consistent AI interactions today.
This repository contains the source for the GitHub Pages site:
awesome-ai-native/
├── docs/ # Main guide sections
│ ├── concepts/ # Engineering principles
│ ├── getting-started/ # Foundation setup
│ ├── workflows/ # Advanced orchestration
│ ├── team-adoption/ # Scaling strategies
│ └── reference/ # Quick lookups
├── _examples/ # Ready-to-use templates
│ ├── instructions/ # Domain-specific guidance
│ ├── chatmodes/ # Role-based AI specialists
│ ├── prompts/ # Workflow templates
│ └── specifications/ # Implementation blueprints
├── index.md # Site homepage
└── _config.yml # Jekyll configuration
We welcome contributions that advance AI Native Development research and practice! Whether you're sharing experimental results, contributing Agent Primitives, or improving documentation—your expertise helps push the boundaries of AI-assisted programming.
📖 Read the Contributing Guide →
This project is licensed under CC BY-NC-SA 4.0, enabling free educational use while supporting sustainable commercial development. Contributors retain copyright to their work and are credited in all derivative applications.
This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
- ✅ Free for education, research, and non-commercial use
- ✅ Attribution required
- ✅ Derivatives must use same license
- ❌ Commercial use requires permission
For commercial licensing inquiries (corporate training, book publishing, etc.), please contact @danielmeppiel.
🌟 Community Resources: Explore the Awesome GitHub Copilot repository for hundreds of community-contributed instructions, prompts, and chat modes across all major languages and frameworks.
