The prose-critique skill provides a rigorous review framework for creative writers using Claude Code. Unlike standard AI feedback that may offer vague praise, this skill adopts an adversarial mindset to actively challenge assumptions and identify what isn't working across five key dimensions: structure, character, voice, prose, and continuity. It includes specialized resources for each focus area and can perform mechanical analysis of sentence variety and dialogue ratios. By calibrating its feedback to the draft's current stage—from early structural foundations to late-stage line polishing—it ensures writers receive actionable, non-obvious findings that improve the final reading experience.
主要功能
01Reasoned and directable findings that link every critique to a concrete cost for the reader.
02Adversarial reading methodology designed to challenge narrative assumptions and find failures.
03Mechanical prose analysis for quantitative signals like sentence length and repetition.
04Targeted focus areas for structure, character development, voice, prose, and continuity.
05Draft-stage calibration to prioritize structural fixes over prose polish in early versions.
06243 GitHub stars
使用场景
01Refining prose rhythm and word choice through mechanical analysis and line-level critique.
02Identifying plot holes and character motivation inconsistencies in an early-stage draft.
03Performing a deep-dive audit of narrative voice and dialogue consistency in a mid-stage manuscript.