slide-excellence

name	slide-excellence
description	Comprehensive slide excellence review combining visual audit, pedagogical review, and proofreading. Produces three reports and a combined summary.
disable-model-invocation	true
argument-hint	[QMD or TEX filename]

Slide Excellence Review

Run a comprehensive multi-dimensional review of lecture slides. Three agents analyze the file independently, then results are synthesized into a combined summary.

Steps

1. Identify the File

Parse $ARGUMENTS for the filename. If no argument provided, ask the user which lecture to review.

Resolve the full path:

• If argument is just a name like Lecture4_StaggeredRollout.qmd, look in Quarto/
• If argument is a .tex file, look in Slides/
• If argument includes a path, use it directly

2. Run Review Agents in Parallel

Launch all three agents as Task subagents simultaneously.

Agent 1: Visual Audit (slide-auditor)

• Check overflow, font consistency, box fatigue, spacing, images
• Format-aware: Quarto-specific suggestions for .qmd, LaTeX-specific for .tex
• Save report to: quality_reports/[FILENAME_WITHOUT_EXT]_visual_audit.md

Agent 2: Pedagogical Review (pedagogy-reviewer)

• Check 12 pedagogical patterns, deck-level narrative, pacing, notation
• PhD student lens: prerequisites, objections, worked examples
• Save report to: quality_reports/[FILENAME_WITHOUT_EXT]_pedagogy_report.md

Agent 3: Proofreading (proofreader)

• Check grammar, typos, consistency, academic quality, citations
• Save report to: quality_reports/[FILENAME_WITHOUT_EXT]_report.md

3. Synthesize Combined Summary

After all three agents complete, read their reports and produce a combined summary:

# Slide Excellence Review: [Filename]
**Date:** [date]

## Overall Quality Score: [EXCELLENT / GOOD / NEEDS WORK / POOR]

**Total Issues:** N
- Critical: X
- Medium: Y
- Low: Z

### By Dimension
| Dimension | Critical | Medium | Low |
|-----------|----------|--------|-----|
| Visual/Layout | | | |
| Pedagogical | | | |
| Proofreading | | | |
| TikZ Freshness | | | |
| Content/Env Parity | | | |

### Critical Issues (Immediate Action Required)
1. **[Dimension] [Slide/Location]** — [Description]. [Recommendation].
2. ...

### Medium Issues (Next Revision)
[Grouped by dimension]

### Reports
- Visual: `quality_reports/[FILE]_visual_audit.md`
- Pedagogy: `quality_reports/[FILE]_pedagogy_report.md`
- Proofreading: `quality_reports/[FILE]_report.md`

### Recommended Next Steps
1. Fix critical issues (prioritize pedagogical violations first)
2. Re-render and verify visual fixes
3. Re-run `/slide-excellence` to confirm improvement

Quality Score Rubric

Score	Critical Issues	Medium Issues	Meaning
Excellent	0-2	0-5	Ready to present
Good	3-5	6-15	Minor refinements needed
Needs Work	6-10	16-30	Significant revision needed
Poor	11+	31+	Major restructuring needed

Important Notes

• This is read-only — no files are edited automatically
• All agents run independently (no dependency between them)
• User decides which recommendations to accept
• Can re-run after fixes to track improvement
• For quick single-dimension checks, use /visual-audit, /pedagogy-review, or /proofread individually

Typical Workflow

/slide-excellence Lecture5_DiD.qmd
# → Review combined summary
# → Fix critical issues
# → Re-run to verify
# → Commit when score is "Good" or "Excellent"