Temporal Agent CLI - examples for AI-native workflow debugging

[moedash]

This PR adds a collection of example applications demonstrating how AI agents can use Temporal's new debugging commands (temporal workflow failures, temporal workflow diagnose, etc.) to build and troubleshoot distributed systems.

What's included

Example applications:

• agent-demo - Simple workflow demonstrating the debug loop pattern
• debug-loop - Order processing workflow with intentional failure scenarios
• ai-research-agent - Multi-step research workflow with parallel activities
• ecommerce-nexus - Multi-namespace e-commerce system using Nexus for cross-namespace calls
• ticketdrop - Ticket sales system with queue management and payment processing

Documentation:

• Step-by-step plans for building each application with an AI assistant
• LLM chat transcripts showing real debugging sessions
• Demo scripts for showcasing AI-assisted debugging in presentations

Why this matters

These examples validate that the new CLI commands provide enough context for AI agents to:

1. Identify failing workflows without manual investigation
2. Trace through child workflows and Nexus operations to find root causes
3. Generate visual diagrams (Mermaid) to understand workflow state
4. Fix bugs based on structured error output

The ticketdrop-impl example includes a demo workload script and a pre-configured bug for live demonstrations of AI-assisted debugging.

[moedash]

Closed to avoid branching on the public repo.

[moedash]

Here's a replacement PR, done on a fork: moedash#1