Add workshop docs, repo hygiene, and notebook CI

[Divergent-Code]

This repo shipped as two notebooks + a LICENSE with no onboarding. This PR adds the scaffolding a workshop repo that people fork actually needs. It touches no notebook files, so it's independent of #6 (the correctness/robustness fixes) and can be reviewed on its own.

What's added

• README.md — what the project is, the V1 → V2 progression, "Open in Colab" badges for each notebook, the GOOGLE_API_KEY prerequisite + Colab Secrets setup, local-run instructions, and a short PII note.
• .gitignore — Jupyter checkpoints, Python caches, virtualenvs, local Chroma persistence, and editor/OS cruft (so students don't accidentally commit them).
• requirements.txt — the notebooks' dependencies for local (non-Colab) runs, with minimum-version pins and a note that Colab preinstalls several.
• .github/workflows/validate-notebooks.yml — CI that validates each notebook's JSON structure and compiles every code cell on push/PR. It is structure/syntax only — it does not execute the notebooks (that would need a Gemini key + heavy ML deps). This mirrors the manual checks used in Fix RAG eval, guardrail ordering, and agent indexing in v2 notebook #6 and stops a broken notebook from ever landing.

Notes

• The README badges point at camunity/main, so they resolve correctly once merged here.
• The CI workflow was tested locally against the current notebooks (both pass).
• Pairs with Fix RAG eval, guardrail ordering, and agent indexing in v2 notebook #6; either can merge first.

🤖 Generated with Claude Code