PenglongHuang/chinese-novelist-skill

What it solves

This project provides a structured framework for AI to write complete, long-form Chinese novels. It addresses the common problem of AI struggling to maintain consistency and persistence when writing extended narratives, ensuring the story is completed from initial concept to final polish.

How it works

The system operates through a four-phase automated pipeline:

1. Initialization (Phase 0): Loads user preferences and detects if there are any interrupted projects to resume.
2. Interactive Questioning (Phase 1): Uses a three-layer progressive questioning system to define the core plot, protagonist, and deep customizations (world-building, perspective, and themes).
3. Planning (Phase 2): Generates a detailed outline, character profiles, and a machine-readable writing plan in JSON format.
4. Automated Writing (Phase 3): Executes the writing process using one of three modes (Serial, Sub-Agent Parallel, or Agent Teams). Each chapter is analyzed, written (3,000-5,000 words), polished to remove "AI-style" phrasing, and checked for word count.
5. Validation (Phase 4): Automatically checks the entire manuscript for word count and coherence, triggering automatic rewrites for failing sections (up to 3 rounds).

Who it’s for

Aspiring authors and AI enthusiasts who want to generate high-quality, long-form Chinese fiction with minimal manual intervention and professional narrative structures.

Highlights

• Preference Memory: Learns user preferences across sessions for faster setup.
• Flexible Writing Modes: Supports serial writing, parallel sub-agents for speed, and multi-agent teams for large-scale works.
• Interruption Recovery: Automatically detects breakpoints to resume writing.
• Professional Writing Guides: Includes built-in references for hook techniques, dialogue writing, and plot structures.
• Automatic Quality Control: Integrated word count scripts and coherence checks to ensure professional standards.