Tomato Novelist: The High-Data Fiction Creator

The Core Problem It Solves

On the Tomato Novel platform, the algorithm's sensitivity to chapter word count and emotion tags directly impacts a work's recommendation traffic. Authors face two key challenges: accurately grasping the content rhythm and emotion curve that align with algorithmic preferences, and maintaining a chapter word count within the 2200-2800 word range to optimize metrics like completion rate. This skill provides a structured creative framework that breaks down "hit factor" elements into executable steps.

How the Skill Works

The assistant employs a three-phase, pipeline-like creative workflow. Its core capabilities and key steps are:

1. Phase 1: Data-Driven Gene Locking

   • Through an interactive 6-question confirmation process (using multiple-choice), it defines the novel's core elements, such as emotion tags (e.g., "face-slapping power fantasy") and core conflict. This transforms a vague idea into parameters for structured processing.

2. Phase 2: Structured Outline Generation

   • Based on the confirmed parameters, it automatically creates the project folder and core documents:
     • 00-Outline.md: Generates the overall chapter plan and TODO list using a template.
     • 02-EmotionCurve.md: Plans the emotional arc and climax points for each chapter.
     • 03-GoldenOpening.md: Outputs multiple versions of the opening 50-word excerpt for selection.

3. Phase 3: Controlled Chapter-by-Chapter Creation & Feedback

   • This is the core execution phase. Each chapter follows a strict pre-writing, writing, and post-writing workflow.
   • Critical Control Point: After each chapter, a word count check (using python scripts/check_chapter_wordcount_tomato.py) is mandatory to ensure it falls within the 2200-2800 word range. A series of polishing rules (e.g., "remove over-adorned adjectives", "add colloquial expressions") are applied to reduce AI-generated artifacts.
   • Data-Driven Adjustment: Every 5 chapters, a phase rhythm report is generated, and user feedback (e.g., follow-read rate) is sought. The subsequent outline is dynamically adjusted based on actual operational data.

Applicable Scope and Caveats

• Platform-Specific: The skill's templates, word count rules, and emotion curve designs are tailored for the Tomato Novel platform's ecosystem and may require adjustment for direct use on other platforms.
• Rigid Word Count Constraint: The per-chapter word count is a hard boundary. The workflow includes the automated check_chapter_wordcount_tomato.py script for enforcement.
• "De-AI-fication" is an Ongoing Process: The skill emphasizes executing specific actions during each polish (e.g., "avoid stacking four-character idioms", "alternate between long and short sentences"), rather than a one-time style conversion.
• Dependency on User Feedback: The "data feedback" step in Phase 3 is crucial for adjusting subsequent outlines. If the user cannot provide actual operational data, the effectiveness of the adjustment phase will be limited.