Replace Chapter 8 with Microsoft Waza eval lab

[Copilot]

Bug Fix

What was the bug?

Chapter 8 relied on third-party eval tooling with licensing assumptions customers may not have. The lab also lacked the Waza workspace artifacts required to evaluate chat customizations.

How did you fix it?

• Chapter 8 content

  • Replaced the prior eval/observability lab with Microsoft Waza guidance.
  • Added GitHub Actions and VS Code Chat Customizations Evaluations workflows.
  • Clarified that customers should use customer-owned eval data.

• Waza CI

  • Added .github/workflows/waza-eval-suite.yml.
  • Runs Waza against evals/token-optimization/eval.yaml.
  • Uploads JSON and JUnit result artifacts.

waza run evals/token-optimization/eval.yaml \
  --verbose \
  --output waza-results/results.json \
  --reporter junit:waza-results/results.xml

• Required customization artifacts

  • Added a workspace skill: .github/skills/token-optimization.skill.md
  • Added a prompt: .github/prompts/context-triage.prompt.md
  • Added instructions: .github/instructions/token-optimization.instructions.md
  • Added an agent: .github/agents/token-eval-reviewer.agent.md

• Waza eval scaffold

  • Added skills/token-optimization/SKILL.md.
  • Added a minimal eval suite under evals/token-optimization/.

Testing

Not included in this description; review should focus on the Chapter 8 replacement, Waza workflow, and required workspace customization files.