GenUI and AI Full-Stack Hub
Use this hub when
- AI generates UI or workflow state
- a model retrieves product data
- an agent calls tools
- MCP/A2A-style integrations enter product architecture
- you need eval gates, trace review, or kill switches
Reading path
| Step | Read |
|---|---|
| Foundation | part-xiii/generative-ui-architecture, part-xiii/new-ai-full-stack-map |
| UI contracts | part-xiii/react-genui-implementation-patterns, part-xii/completed-examples/completed-genui-registry-contract |
| Retrieval/tools | part-xiii/rag-vector-data-for-ui, part-xiii/mcp-a2a-integration-boundaries |
| Production | part-xiii/production-genui-playbook, part-xiii/genui-evaluation-governance |
| Examples | part-xii/completed-examples/completed-genui-launch-review, part-xii/completed-examples/completed-genui-eval-matrix |
| Capstone | capstones/capstone-genui-ai-fullstack-copilot, capstones/solution-packs/capstone-solution-pack-genui-copilot |
Artifacts to produce
- workflow boundary diagram
- component registry contract
- eval matrix
- trace schema
- tool risk matrix
- kill-switch matrix
- launch review
Review prompts
- What product authority remains deterministic?
- What can the model propose but not execute?
- Which retrieval sources are permission-filtered before context assembly?
- Which eval failures block launch?
- Which kill switch disables each layer?
Source lens
Use current official AI/protocol docs for implementation details and Part XIII for stable architecture controls around UI, tools, evals, observability, and governance.