docs(code-review): add experiment leaderboard table

[gggdttt]

Adds the Experiment Leaderboard section to the code-review docs page, decoupled from #696.

Why

#696 (live BCQuality consumption) is paused pending a review of how BCApps consumes BCQuality, but we still run the experiment data. The site page should reflect those runs independently of merging #696.

Changes

• docs/code-review.md: rename baseline header F1 -> Micro F1 for consistency; add an Experiment Leaderboard table that renders rows where agg.experiment is set (BCQuality live skills vs inline pre-#8700 knowledge).
• src/bcbench/types.py: add bcquality: bool = False to ExperimentConfiguration (and include it in is_empty()). Required so the leaderboard aggregation preserves the live-arm marker; otherwise pydantic drops the unknown key and live-arm rows mislabel as Other. Default-off and backward compatible — existing results are unaffected.

Safety

• Self-contained: with no experiment data the table shows No experiment results available yet. No pipeline / value-computation changes; gracefully degrades. The oldinline arm already labels correctly on main; this PR additionally enables the BCQuality live-arm label.

Generated with the help of GitHub Copilot.