Experiment: composed BCQuality super-skill/sub-skill code review

[WaelAbuSeada]

Adds a new code-review experiment that implements BCQuality's composed super-skill / sub-skill review pattern.

What this adds

• Vendors the BCQuality composed-review framework into src/bcbench/agent/shared/instructions/microsoft-BCApps/skills/al-code-review/ as committed static copies (no runtime fetch).
  • Meta-skill contracts (read.md, do.md)
  • Super-skill al-code-review.md + 5 domain leaf sub-skills (performance, security, privacy, upgrade, style)
  • 123 knowledge articles + AL samples across the 5 domains, plus a generated knowledge-index.json
  • UI domain dropped to match BC-Bench's 5 domains
• Authored Copilot SKILL.md entry point that orchestrates the composition and maps BCQuality findings to BC-Bench's review.json schema (blocker->critical, major->high, minor->medium, info->low; from-sub-skill->domain).
• config.yaml: code-review-template invokes the skill directly via /al-code-review (matching the experiment/code-review-al-skill pattern); skills.enabled: true.

Notes

• CI's existing setup_agent_skills copies the committed skill folder into repo/.github/skills/ — no BCQuality clone needed.
• Relevant tests pass (agent-skills, experiment-config, copilot-prompt, dataset-integrity).