Apply cuts
Apply adversarial-edit cuts to chapter files (mechanical quote-matching removal). Wraps scripts/apply_cuts.py. Filters by cut type (OVER-EXPLAIN, REDUNDANT typically yield ~55-60% of cuts). Use after adversarial-edit during revision cycles.From its SKILL.md
npx -y skills add tchr-dev/autonovel --skill apply-cutsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
- runs commandsInstructs the agent to run 4 commands, including `AUTONOVEL_NOVEL_DIR=<novel-dir> python scripts/apply_cuts.py all --dry-run` and 3 more.
SKILL.md
1.8 KB, 386 tokens by cl100k_base, as published. Nobody here has run it
Apply cuts
Mechanical pass — no LLM call. Reads <novel-dir>/edit_logs/chNN_cuts.json produced by adversarial-edit and removes matching quotes from chapters.
Common invocations
# Dry run on all chapters with cuts files
AUTONOVEL_NOVEL_DIR=<novel-dir> python scripts/apply_cuts.py all --dry-run
# Apply only OVER-EXPLAIN and REDUNDANT cuts (the safest/most common types)
AUTONOVEL_NOVEL_DIR=<novel-dir> python scripts/apply_cuts.py all --types OVER-EXPLAIN REDUNDANT
# Apply all cuts to chapters where adversarial-edit found ≥17% fat
AUTONOVEL_NOVEL_DIR=<novel-dir> python scripts/apply_cuts.py all --min-fat 17
# Single chapter
AUTONOVEL_NOVEL_DIR=<novel-dir> python scripts/apply_cuts.py 12
Failure modes
not found— the quote doesn't appear in the chapter (often because a previous cut already shifted whitespace; the script tries whitespace-normalised match before giving up)ambiguous (N matches)— the quote appears more than once; the script refuses to guess which instance was meant- Skipped if quote < 25 chars
The script collapses runs of 3+ newlines down to 2 after applying cuts.
After running
Report:
- Total words removed across chapters
- Per-chapter applied / failed / skipped breakdown
- Which chapters had the highest cut counts (revision priorities)
Suggest:
- Re-evaluate any heavily-cut chapter with
evaluate-chapterto confirm the score moved in the right direction - If
failedcount is high, manually inspect the cuts file — some quotes may need rephrasing in the cuts JSON before retrying
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.