---
description: Generate structural and statistical analysis of a manuscript
argument-hint: "[path, glob, or selection]"
allowed-tools: [Read, Glob, Grep]
---

Analyze the manuscript at `$ARGUMENTS` (a file path, a glob of chapter files, or pasted text). If no argument is given, use the current selection or the most recently opened manuscript file.

Produce a structural and statistical report. Cite specific numbers — avoid vague assessments.

## Metrics to compute

- **Word and scene counts**: total words, chapter count, scene count, average lengths, longest/shortest chapter and scene
- **POV distribution**: who narrates, how often, and how POV is balanced across chapters
- **Character usage**: appearance frequency and dialogue volume for named characters; flag overload (too many minor characters) or imbalance (one character dominating)
- **Dialogue vs. narrative ratio**: rough percentage, and whether the balance shifts across the book
- **Pacing and scene shape**: scene-length variation, action vs. reflection balance, chapter-ending types (cliffhanger, resolution, transition)
- **Plot and conflict mapping**: active vs. resolved plotlines, internal vs. external conflicts, where tension peaks and troughs

## Output format

Present a compact Markdown report with these sections:

1. **Overview** — headline metrics (words, chapters, scenes, completion status if discernible)
2. **Characters** — frequency table, POV distribution, dialogue volume
3. **Plot and pacing** — plotline status, scene-type distribution, tension curve commentary
4. **Issues and observations** — patterns worth surfacing (imbalances, gaps, structural concerns)
5. **Recommendations** — three to five concrete next steps the writer could take

Output directly to the conversation. If the user wants the report saved, they will ask.