Add FY26-H2 AgenticOdyssey workshop content: Labs and Demos

[Copilot]

Bootstraps the AgenticOdyssey workshop repository with all hands-on labs, live demos, and supporting infrastructure for the FY26-H2 workshop on building AI agents with Azure OpenAI and AutoGen.

Labs (labs/)

• Lab 01 — ReAct Agent: Minimal function-calling agent loop (Thought → Action → Observation) using Azure OpenAI; participants add tools and trace multi-step reasoning
• Lab 02 — Assistants API: Persistent-thread research assistant with code interpreter and web_search function calling; includes starter with a NotImplementedError stub to wire real Bing Search
• Lab 03 — Multi-Agent (AutoGen): Four-agent GroupChat (Planner / Researcher / Coder / Critic); starter has TODO system prompts and asks participants to add a Documenter agent
• Lab 04 — RAG Agent: Document ingestion into Azure AI Search with vector embeddings (ingest.py), then a hybrid-search tool wired into an agent loop; starter upgrades keyword search to hybrid

Each lab ships solution/ + starter/ with guided TODOs, and a README.md covering objectives, step-by-step instructions, and expected output.

Demos (demos/)

• Demo 01 — Coding Agent: AutoGen AssistantAgent + UserProxyAgent loop that writes code, runs pytest, reads failures, and self-patches until tests pass
• Demo 02 — Research Pipeline: Five-agent pipeline (Orchestrator → Researcher → FactChecker → Writer → Editor) that extracts a structured Markdown brief from <brief>…</brief> tags and saves it to disk

Supporting files

• docs/SETUP.md — Azure resource provisioning walkthrough + local env setup
• docs/validate_setup.py — runnable preflight check (Python version, packages, .env keys, live Azure OpenAI ping)
• requirements.txt, .env.example, .gitignore, LICENSE, CONTRIBUTING.md

Original prompt

This repository is to collaborate and build the contents for the FY26-H2-AgenticOdyssey workshop. Labs and Demos

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.