Qimao Novelist Original Writing Assistant

Challenges Posed by the Platform's Model

The Qimao novel platform operates on a "free-to-read + inter-chapter ads + coin incentives" model. This cultivates a reader base that is highly utilitarian, impatient, and motivated by immediate gratification rather than slow-burn plotting. Author revenue is directly tied to total ad impressions (reading duration) and user acquisition via social sharing. Traditional "slow build" and "foreshadowing" approaches perform poorly under these constraints. Therefore, writing must revolve around a set of quantifiable data levers, not just literary merit.

Core Capabilities and Workflow

This skill offers an engineering-driven writing solution tailored to the platform's logic. Its core workflow is divided into three phases:

1. Platform Logic Internalization and Interactive Planning: Before creation begins, the skill enforces user comprehension of Qimao's "Five Levers" (Ad Retention, Read-Through Rate, Social Sharing, Full Attendance Contract, Adaptation Potential). It then uses a two-round interactive selection process (via the AskUserQuestion tool) to collaboratively define the novel's emotional tag, genre, protagonist's reversal point, core conflict, and chapter plan. All selections are recorded in meta.md as the project's single source of truth, aligning from the start with the platform's hit formula.

2. Rule-Based, Process-Driven Creation: After planning, the skill enters "Frenzy Creation" mode. It strictly follows a "must-execute-per-chapter" process:

   • Pre-Write Check: Automatically loads the outline, chapter plan, and the character/location/plot memory bank (.learnings/ directory) to ensure narrative continuity and prevent plot holes.
   • Structured Writing: Mandates a conflict within the first 150 words. Each chapter must contain a "Bully-Counter Punch-Suspense" satisfaction loop. Chapters end with a cliffhanger "you need to know even if you have to sit through the ad" and a "shareable" high-impact quote. Chapter length is strictly controlled between 2,200-3,000 words to match ad intervals.
   • Data Feedback & Adjustment: Every 5 chapters, it generates a pacing report and simulates user data feedback (reading duration, read-through rate, etc.), adjusting the subsequent outline and character focus accordingly.
   • Engineering-Grade QA: Includes a 9-dimension review mechanism that, beyond standard checks, specifically evaluates shareable quote density, continuation hook strength, and ad-retention friendliness—Qimao-specific metrics. Automated scripts (auto_backfill_qimao.py) backfill quality data into the outline.

3. Long-Form Management and Memory Maintenance: For serials over 50 chapters, the skill employs volume-based management (every 50 chapters is one volume) and builds a project memory system (.learnings/ directory), recording all changes to characters, locations, and plot threads. This prevents issues like dead characters reappearing or plot contradictions in long-form writing.

Scope and Considerations

This is not a general-purpose writing assistant; its design is strictly targeted at Qimao's traffic monetization logic. It is especially suited for creating emotional, high-intensity male/female-oriented power fantasies (e.g., Son-in-law, War God, Doting Family, Rebirth) and long-form serials requiring stable daily updates for full attendance bonuses.

Users must understand and accept its built-in writing formulas. The skill's highly process-oriented nature means it functions best as a "collaborator" rather than an "independent creator." Authors still need to intervene for final plot decisions, emotional depth, and literary polish, especially when handling non-formulaic plots or deep character arcs. It cannot replace an author's unique creativity and observational insight.