scan-recurrenceslisted

# Skill: scan-recurrences **What it does.** Sweeps the entire book for phrases, images, and structures appearing in multiple chapters. Distinguishes **intentional recurrence** (a motif stitching the work together) from **accidental duplication** (the same aphorism recycled). Updates `.project/reports/recurrences.md`. **Triggers** - "run scan-recurrences" - "/scan-recurrences" - "find duplications in the book" - "what repeats across chapters?" - Equivalent phrasing in the project's output language **Optional inputs** - Scope: whole book (default), a single Act, or a chapter list - Mode: `full` (complete sweep) or `incremental` (only new/modified chapters) **Output** - Updated `.project/reports/recurrences.md` - Chat report of new duplications > **Write all output in the project's output language.** --- ## Why this skill exists Chapter-by-chapter analysis is blind to the whole book. It can see that a serial comparison is a tic — but not that the same sentence appears in three different chapters. **Literal cross-chapter duplication is the most damaging marker for a critical reader**, because it has no stylistic defence. A human author may have tics; they rarely repeat the same striking sentence three times without noticing. This skill is the only way to catch it. --- ## Execution protocol ### Step 1 — Load context 1. **`.project/config/project.yaml`** — output language 2. **`.project/reports/recurrences.md`** — current state (for incremental mode and to preserve de