Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/epicsagas-book-edit
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/book-edit
What's inside
- Skill name
book-edit- Description
- Phase 4: Editing. Fact-check (nonfiction), assessment, developmental edit, line edit, copy edit, proofread, then a cold read by simulated readers that issues the readiness verdict. Rewrites the manuscript; does not only file reports.
- License
- Apache-2.0
- Instructions
- 91 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Phase 4: Editing
Goal: a manuscript that passes a cold read by its target readers. Editing here means rewriting. Each stage snapshots the drafts, changes them, and files a report of what changed and why.
Read `${CLAUDE_PLUGIN_ROOT}/skills/loom/quality-bar.md`. Every stage reads the whole manuscript, `STYLE.md` (voice lock passages), `bible.md`, and the genre skill.
Refuse to run if any outlined chapter is missing. Editing an incomplete book fixes the wrong things.
## Stages
| # | Stage | Who | Rewrites? | Report |
|---|-------|-----|-----------|--------|
| 0 | Fact check (nonfiction, technical, academic, and fiction with real-world claims) | `fact-checker` | Yes, removes or corrects unverifiable claims | `edits/00-fact-check.md` |
| 1 | Editorial assessment | skill (you) | No | `edits/01-assessment.md` |
| 2 | Developmental edit | skill + `chapter-writer` for rewrites, `continuity-editor` after | Yes, chapter-level | `edits/02-developmental.md` |
| 3 | Line edit | `style-doctor` then skill | Yes, paragraph and sentence level | `edits/03-line-edit.md` |
| 4 | Copy edit | skill | Yes, mechanical | `edits/04-copy-edit.md` |
| 5 | Proofread | skill | Yes, final | `edits/05-proofread.md` |
| 6 | Readiness | `beta-reader` | No | `edits/06-readiness-report.md` (also `edits/readiness-report.md`) |
Before every rewriting stage: `node ${CLAUDE_PLUGIN_ROOT}/velith.mjs snapshot [project-dir] stage{N}`.
After each stage: update chapter frontmatter `status: edited`, bump `revision`, run `scan`.
### Stage 0 — Fact check
Invoke `fact-checker`. It builds a claim ledger, verifies against `sources/` and the web, and edits drafts to remove or correct what cannot be verified. It also reconciles every figure's labels and plotted values with the text and the data files in `sources/`. A claim that would embarrass the author in a review is a Critical. Fiction runs this stage only if the bible flags real-world claims (real places, real history, real science). Stage 0 runs before anything else because a developmental edit built on a false claim is wasted.
### Stage 1 — Editorial assessment
Read the whole book once at reading speed (cold-read protocol). Produce the macro diagnosis:
- Where the book delivers its promise and where it does not (against `PRD.md` reader promise)
- Pacing map: per chapter, engagement 1-5 and why; identify the sag
- Structural redundancies and gaps (two chapters doing one job; a job nobody does)
- Proportion: is the setup too long, the ending too fast
- Voice consistency across chapters (quote a line from chapter 2 and one from chapter 18 side by side)
- Five-axis score for the whole manuscript
- Ordered list of developmental interventions with expected impact
No rewriting in this stage. The report is the plan for stage 2.
### Stage 2 — Developmental edit
Execute the stage 1 plan. Interventions, in order of scale:
- **Restructure** (reorder, merge, split, cut chapters): requires author approval. Present the proposal with rationale and the concrete new chapter list. Required stop.
- **Rewrite chapter** (purpose unchanged, execution replaced): invoke `chapter-writer` with the assessment findings for that chapter as extra input and `--force`. It reads the whole book, so it will match the voice lock.
- **Add or remove scenes/sections** within a chapter: do it directly, in the voice, then update the bible ledger.
- **Deepen** (a character who acts without wanting; an argument missing its strongest objection): targeted rewrites of the passages, not the chapter.
After all interventions, invoke `continuity-editor` on the full manuscript. Report: what changed per chapter, what was left as-is and why, and the new pacing map.
### Stage 3 — Line edit
Invoke `style-doctor`. It measures (`velith.mjs metrics`), reads, rewrites paragraphs and sentences in place, and reports before/after with metrics. Then do your own line pass on anything it flagged as "author decision": sentences whose problem is meaning, not style.
Line-edit priorities: rhythm uniformity, the structural tells, emotional tells, dialogue that answers, figurative density against the fingerprint, paragraphs that end on a punch, Korean 번역투 and ~것이다 endings, ~5% trim of every chapter by cutting what the reader already knows.
### Stage 4 — Copy edit
Mechanical consistency against `bible.md` and `STYLE.md`: spelling of names and terms, hyphenation, number style, capitalization, quotation marks, 띄어쓰기, tense agreement, heading levels, code formatting (technical), citation format (academic), cross-references that point to the right chapter, figure numbering and caption format, every figure referenced in the text before it appears. Run `node ${CLAUDE_PLUGIN_ROOT}/velith.mjs images check [project-dir]` and fix unresolved references and missing alt text. Fix directly. Report only counts by category plus anything ambiguous for the author.
### Stage 5 — Proofread
Final pass on the revised text: typos, doubled words, orphaned formatting, markdown artifacts, missing scene breaks, dialogue punctuation, image references that resolve. Fix directly. Then set every chapter `status: final`.
### Stage 6 — Readiness
Invoke `beta-reader`. It cold-reads the whole manuscript as three target readers and one genre expert, writes `edits/06-readiness-report.md` with per-chapter engagement, put-down points, confusion log, AI-feel log with quotes, comp comparison, five-axis scores, and a verdict.
- `verdict: PASS` — write `edits/editorial-report.md` summarizing all stages and the scores. Phase complete.
- `verdict: REVISE` — the report includes a prioritized fix list. Apply the fixes (stage 2 or 3 interventions as appropriate), then rerun stage 6. Maximum two loops, then present the report to the author with the remaining issues and ask whether to override or continue. Required stop.
Copy the final report to `edits/readiness-report.md` (the router and dashboard read this path).
## Running a single stage or chapter
`/book-edit 3` runs only the line edit. `/book-edit 3 drafts/ch07-x.md` line-edits one chapter (still with full-book context). Snapshots and reports are still written.
## Gate
`edits/readiness-report.md` has `verdict: PASS` (or an author override recorded in `edits/editorial-report.md`). All chapters `status: final`. Stage reports 01-06 exist (00 for nonfiction). No unresolved Critical issues in the last continuity or fact-check report.
## Post-completion
```bash
node ${CLAUDE_PLUGIN_ROOT}/velith.mjs scan [project-dir]
```From epicsagas/Velith, Apache-2.0.
More from epicsagas/Velith
- 0.641score
book-visuals
epicsagas
Book visual system: art bible and look lock, figure plan per chapter, code-rendered diagrams/charts/technical drawings, diffusion-generated illustrations and…
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-draft
epicsagas
Phase 3: Drafting.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-fiction
epicsagas
Fiction craft reference for all subgenres: structure options, scene design, character depth, dialogue, POV and distance, the fiction-specific AI tells, and…
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-game
epicsagas
Game narrative craft reference: branching architecture, quest design, dialogue systems, lore bible, player agency, and game-writing AI tells.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-genre-creator
epicsagas
Meta-skill for genre selection and custom genre creation.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-ideation
epicsagas
Phase 1: Ideation.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-nonfiction
epicsagas
Nonfiction craft reference: argument architecture, evidence discipline, narrative in nonfiction, chapter design, the nonfiction-specific AI tells.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-outline
epicsagas
Phase 2: Outlining.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-poetry
epicsagas
Poetry craft reference: forms, the line, image, sound, compression, collection architecture, and poetry-specific AI tells.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-publish
epicsagas
Phase 5: Publishing.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-screenplay
epicsagas
Screenplay craft reference: format, three-act and sequence structure, scene construction, dialogue subtext, A/B/C story weaving, series bibles, and…
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
loom
epicsagas
Velith router.
SkillAFreshness: A★ 38Claude Code +9
Signals
- Stars
- 38
- Forks
- 6
- Open issues
- 1
- Last commit (this folder)
- Sep 4, 2026 (1 month ago)
- Last release
- Sep 30, 2026 (28 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/epicsagas/Velith
- Website
- epicsagas.github.io/Velith/
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from epicsagas
- 0.641score
book-visuals
epicsagas
Book visual system: art bible and look lock, figure plan per chapter, code-rendered diagrams/charts/technical drawings, diffusion-generated illustrations and…
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-genre-creator
epicsagas
Meta-skill for genre selection and custom genre creation.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-academic
epicsagas
Academic writing craft reference: IMRAD and discipline structures, literature review, argument chains, citation integrity, academic register in en/ko/ja, and…
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-draft
epicsagas
Phase 3: Drafting.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-game
epicsagas
Game narrative craft reference: branching architecture, quest design, dialogue systems, lore bible, player agency, and game-writing AI tells.
SkillAFreshness: A★ 38Claude Code +9 - 0.638score
book-fiction
epicsagas
Fiction craft reference for all subgenres: structure options, scene design, character depth, dialogue, POV and distance, the fiction-specific AI tells, and…
SkillAFreshness: A★ 38Claude Code +9
Cite this page
book-edit skill. Agent Mart. https://agentmart.dev/skills/epicsagas-book-edit (last updated 2026-10-24)