AgentKit Added Prompt, Parser, Memory, RAG, and Workflow Examples

Update Notes

AgentKit expanded its demo coverage around chain-style and graph-style AI application patterns, using Qt/C++ APIs while referencing common agent development workflows.

This update includes:

• Added prompt templates, structured output parsers, chat memory, runnable chains, and parallel execution.
• Added in-memory retrieval and simple RAG flow for document-driven answers.
• Added state workflow support with nodes, conditional routes, loops, checkpoints, interrupts, and graph traces.
• Added independent demo pages so each capability can be tested and explained separately.

These examples make AgentKit easier to evaluate as a product toolkit rather than only a single chat demo.