Fanqie Novel Anti-AI Writing Assistant

Problem

This skill targets a practical writing problem: long-form web novels for Fanqie often become too generic, too formulaic, or too AI-sounding. When using an LLM for genre fiction, writers may get polished but hollow prose, identical character voices, drifting worldbuilding rules, and chapters that lack the fast conflict hooks expected by mobile readers.

How It Works

It structures the process into six stages: genre positioning, worldbuilding, character design, volume outline, chapter drafting, and serialization optimization. It uses templates and reference files to constrain outputs, such as references/genre_guides.md for topic selection, assets/worldbuilding_template.md for rules and limits, assets/character_profile_template.md for voice and state tracking, and references/anti_ai_writing.md for anti-AI phrasing.

• Anti-AI style: blocks common filler phrases, enforces varied sentence length, colloquial dialogue, and action-based interiority.
• Long-form consistency: tracks character power, location, relationships, goals, and unresolved foreshadowing before drafting.
• Platform pacing: favors early conflict, short paragraphs, separate dialogue blocks, chapter hooks, and 2000-3000 word chapters.

It is best for fast-paced male or female audience web fiction and for turning model output into platform-ready draft rhythm. It is less suited to slow, literary, or experimental writing unless the strong conflict and hook requirements are softened.