webnovel-review_skill

Overview

This skill reviews webnovel chapters using multiple checker agents and generates structured quality reports. It runs from the project root, loads minimal reference material by review depth, runs parallel checkers, and outputs both human-readable reports and JSON metrics for trend tracking. Use it to find continuity, pacing, characterization, and reader-pull issues across chapter ranges.

How this skill works

First confirm you are in the project root (requires .webnovel/state.json) or prompt for the project path. Determine review depth (Core or Full) and load only the needed reference files. Run a set of checker agents in parallel (core checkers plus optional full-checkers) and aggregate findings into a markdown report and a metrics JSON. Save metrics with the project index manager and prompt the user for immediate fixes if critical issues appear.

When to use it

• When you want a quick automated quality scan of one or more chapters (Core depth).
• When reviewing a key chapter or fixing pacing/high-point problems (Full depth).
• Before publishing a chapter to catch continuity and characterization drift.
• When tracking quality trends across serialized chapters via metrics.
• When you need prioritized, actionable fixes rather than vague feedback.

Best practices

• Run from the project root so .webnovel/state.json is accessible; otherwise change to the correct directory first.
• Start with Core depth for routine checks and escalate to Full for critical chapters or pacing issues.
• Load additional reference files only when the issue requires deeper context to reduce overhead.
• Save the generated metrics JSON for longitudinal analysis and use the index manager to persist them.
• If a critical issue is detected, prefer immediate repair to prevent compounding contradictions.

Example use cases

• Run /webnovel-review on chapters 12-14 before publication to catch continuity errors and weak reader-pull.
• Perform a Full review on a climax chapter to verify high-point, pacing, and emotional beats.
• Batch-scan the last 10 chapters to produce trend metrics and identify recurring issues.
• Generate prioritized fixes after beta reader feedback that mentions character OOC or plot holes.
• Save review metrics to feed a dashboard tracking quality scores across a long serialization.