CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

inkos-story-review

按题材、目标读者和用户标准审稿,展示具体问题并协作修订。Use for chapter or manuscript review with transparent criteria.

SKILL.md
Quality
Evals
Security

Story review

Use this skill when the user wants diagnosis, comparison, review, or revision advice for existing prose.

  • First identify the applicable standard from genre, target audience, tone, platform, and explicit user preference. Everyday comedy, literary fiction, romance, mystery, and commercial serials should not share one logic-density threshold.
  • In an active book, use the auditor for persisted chapters. Show concrete issue descriptions, severity, evidence, and likely reader impact.
  • A parser or model-format failure produces a review-unavailable observation. It is not evidence that the prose is bad and never authorizes a rewrite.
  • Revise when the user asks or an explicit production action requests revision; report whether the revised artifact was actually applied.
  • Preserve the user's voice and successful passages. Prefer the smallest scope that resolves the real defect.
  • For long-form work, compare the chapter with populated memo requirements, recent progression, real hook ids, character knowledge, and reader promises. Repetition is not progression; a payoff, subplot, or emotional line moves only when facts, action, relationship, status, or understanding changes.
  • Route repairs by cause: local prose defects may be patched, while missing scenes, broken causality, timeline, viewpoint, character logic, or payoff require structural revision.
  • Respond in the user's language.

Load references/review-matrix.md for a full review or when standards are disputed. The matrix is qualitative guidance, not a scorecard. For spinoffs, use parent canon as authority; for fanfic, honor fanficMode and explicit allowed deviations.

Repository

Narcooo/inkos

Path

packages/core/skills/inkos-story-review/SKILL.md

Last updated
First committed

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.