feat(report): group categories into Pixel Accuracy / Token Efficiency

[let-sunny]

Summary

• Replace 6 individual category gauges with 2 group cards: Pixel Accuracy (pixel-critical, responsive-critical) and Token Efficiency (code-quality, token-management, semantic, interaction)
• Each group card shows average score, description, and clickable category bar charts
• Mobile: hide bar graphs, show compact inline view (category name + percentage) while keeping 2-column layout
• Display-only change — no scoring formula modification

Motivation

Closes #215 — users see Grade D (51%) despite 95% pixel similarity because token-related rules dominate. Grouping makes it immediately clear: "pixel accuracy is fine, token efficiency is dragging the grade."

Test plan

• 34 render tests passing (new tests for renderCategoryGroup)
• tsc --noEmit passes
• Visual check: desktop shows bar chart layout
• Visual check: mobile (<600px) shows compact inline layout with 2-column grid

https://claude.ai/code/session_01H2EPyD3PhqkQFrk5BB4M4g

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: b1511e7c-d3f6-4e37-927c-fa6c8ca9fd20

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/improve-report-ftTS6

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[let-sunny]

Closing — the underlying problem (D/F grades for well-structured files) was resolved by sqrt damping (#227) instead of display-level grouping. With damping applied, per-category scores are reasonable and grouping is unnecessary.

---

Generated by Claude Code