| name | book-review |
| description | Run multi-persona editorial reviews on a drafted chapter. Seven personas (Robert Gottlieb, Lay Reader, Domain Expert, Copyeditor, Enjoyment Reader, AI-Slop Detector, First-Time Visitor) read the chapter and return severity-tagged feedback. Critical findings soft-gate chapter release. Use when user says "review chapter X with personas", "Gottlieb pass on this chapter", "run the editorial reviews", "what would Gottlieb say about this draft", "is this chapter ready for review", "soft-gate this chapter". Do NOT use for source ingestion (use book-knowledge), prose-only style fixes (use russellian-style), chapter drafting (use book-compose), or persona reviews on prose outside the book pipeline. |
| license | MIT |
| metadata | {"author":"charles-hoskinson","version":"0.1.0","category":"writing","workspace-aware":true} |
book-review
Qualitative editorial review for chapter drafts. Dispatches up to seven persona subagents in parallel, aggregates severity-tagged findings, soft-gates release on persona_critical_count == 0. Panel composition is set per workspace; see review-conductor/panels/.
What it owns
- Persona definitions in
personas/*.md and conventions for new ones.
- Dispatch-packet construction, subagent prompting, review parsing.
- Aggregation into
<workspace>/chapters/drafts/<chapter_id>/persona-review.md.
- Soft-gate semantics: critical blocks; important and minor are advisory.
What it does NOT own
- Mechanical defect-gating on built artefacts — owned by
book-qa.
- Sentence-grain prose discipline — owned by
russellian-style.
- Claim verification, wiki synthesis, SHACL gates — owned by
book-knowledge.
- Chapter drafting, release bundles, book builds — owned by
book-compose.
- Persona rewrites. Personas comment; revisions go back through
book-compose.
Personas
- Robert Gottlieb (
personas/gottlieb.md) — voice, cadence, AI-sloppy patterns.
- Lay Reader (
personas/lay-reader.md) — accessibility, unexplained jumps.
- Domain Expert (
personas/domain-expert.md) — factual accuracy against the claim ledger.
- Copyeditor (
personas/copyeditor.md) — cross-chapter consistency, mechanics.
- Enjoyment Reader (
personas/enjoyment-reader.md) — momentum, where the reader stops.
- AI-Slop Detector (
personas/ai-slop-detector.md) — 24-pattern Wikipedia catalog of AI-writing tells.
- First-Time Visitor (
personas/first-time-visitor.md) — 30-second drive-by; does the opening earn the read?
Severity rubric
A finding is critical iff the persona believes the chapter must not ship as written.
- Critical — blocks release. Reserved for patterns each persona marks as critical. Prefer
important if uncertain.
- Important — fix before publication; counted but non-gating.
- Minor — advisory polish.
The aggregator takes severity at face value. See references/severity-rubric.md for per-persona patterns and dispatch order.
Workflow
- Prepare. Verify the draft exists; load the chapter contract.
- Dispatch.
scripts/review_pass.py:prepare_dispatch_packets returns one packet per persona. Issue parallel Task-tool calls; each subagent writes to packet.output_path.
- Aggregate.
scripts/aggregate_reviews.py:aggregate_reviews merges reports into persona-review.md with counts, verdict table, substring-deduplicated findings.
- Soft-gate. If
critical > 0, surface the critical findings and stop. Otherwise surface important and minor as advisory.
Composes with
book-compose — invokes this skill in Stage 4; gates on persona_critical_count == 0.russellian-style — runs before review so personas see compliant prose.book-knowledge — Domain Expert reads the claim ledger and raw sources.book-qa — runs after on the built manuscript; orthogonal mechanical scope.
Usage
- "Review chapter X with personas" — full seven-persona pass.
- "Gottlieb pass on this chapter" — single-persona dispatch.
- "Is chapter X ready for review?" — preflight check.
Tests
Five suites in tests/: test_persona_loader, test_dispatch_review, test_aggregate_reviews, test_review_pass, test_sibling_skills. Run pytest from the skill root. Synthetic reports in tests/fixtures/synthetic_reviews/.
See also
review-revise-validate — to take a panel's findings and drive a revision cycle through them, use the sibling review-revise-validate skill. It composes this skill's panel + aggregate with a gemma4:31b reviser persona, applies paragraph-level rewrites, re-runs the panel, and reports the before/after delta.
