novel-master for Openclaw

A publication-grade Chinese long-form novel creation framework powered by a unique distributed multi-agent architecture.

bbroot
v1.0.0
Jul 8, 2026
0
395
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install novel-master-test-2026

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install novel-master-test-2026 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is novel-master?

novel-master is an advanced framework engineered specifically for generating professional, publication-quality Chinese fiction. Moving past typical AI-generated tropes and structural loose ends, this framework delivers condensed prose, deep emotional resonance, and precise plot execution capable of meeting strict editorial guidelines for both online web serialization and physical publishing houses.

At its core, version 2 introduces a revolutionary distributed creation architecture that breaks away from monolithic AI memory limitations. By utilizing isolated character subprocesses, fact-compression graphs, and multi-dimensional quality gates, it empowers AI coding agents to comfortably manage complex epics scaling up to millions of words without suffering from continuity breaks or out-of-character behavior.

novel-master Use Cases

  • Drafting high-quality Chinese novels, serialized web-fictions, or short stories from minimal core prompts.
  • Constructing exhaustive world-building assets, geographical maps, character voice profiles, and intricate multi-arc structural outlines.
  • Writing individual chapters under a strict single-character point of view with automated self-checking parameters.
  • Managing complex foreshadowing loops and structural conflict registers over extended narratives.
  • Mitigating standard AI homogenization by introducing targeted cognitive biases and asynchronous plot communication.

How novel-master Works

  1. Book Initiation: The system ingests user-provided inputs like genre, core themes, main character profiles, and target length, cross-references internal architectural standards, and exports a mandatory, highly-structured Creation Strategy Proposal.
  2. Outline Planning: Generates extensive multi-arc narrative macro-outlines and granular chapter breakdowns detailing structural shifts, specific character viewpoints, and explicit hooks.
  3. Chapter Composition: The Chapter Daemon dynamically pulls individual character cognitive packages, style standards, and active foreshadowing registers to write the text following explicit point-of-view limits.
  4. State Broadcaster: Following the completion of a chapter, the Chapter Daemon broadcasts factual events to an internal JSON Plot Bus, updating corresponding character state machines.
  5. Graph Settlement: The Settler component extracts highly compressed factual data nodes (under 200 words per chapter) and populates a Node Compressed Graph (NCG) to act as long-term differential memory.
  6. Quantum Audit Inspection: The Auditor evaluates the submission across a strict four-dimensional boundary (Name, Time, Geography, Foreshadowing) and blocks pipeline progression if conflicts are triggered.

novel-master Setup

The novel-master toolkit operates using integrated python workflows within your agent workspace.

Initialize character processes and run validation audits natively through your CLI:

# Initialize a specific character subprocess with an optional cognitive bias package
python scripts/char_state.py init <book_name> <character_name> [--bias <id>]

# Audit a chapter immediately against four-dimensional logical constraints
python scripts/auditor.py gate <book_name> <chapter_number>

novel-master Data Schema & Taxonomy

Every novel development workspace is meticulously organized under the standard user home path (~/.qclaw/workspace/novels/<book_name>/).

Directory Layout

Path Component File/Folder Type Description
settings/ Directory Houses global configuration, including world.md (worldview parameters), geography.md (factions and maps), rules.md (internal mechanics), and individual files within characters/ (voice cards).
outline/ Directory Contains structural planning blueprints: arc-outline.md (multi-volume arcs) and chapter-outline.md (chapter beats).
chapters/ Directory Holds the raw Markdown textual outputs for completed novel chapters.
tracker/ Directory Active runtime logs including foreshadowing.json (unresolved plants/payoffs), conflicts.json (clash records), and style-log.md (linguistic tracking).
state.json File A high-level status dictionary detailing real-time word counts, the active chapter, unresolved hooks, and metadata.

novel-master Advanced Features

  • Process-per-Character (PpC): Runs every vital character as an independent state machine communicating exclusively through a centralized event bus to prevent identity overlap.
  • Node Compressed Graph (NCG): Condenses every chapter into compressed fact vectors to build lightweight, multi-dimensional timeline maps.
  • Crash Recovery Protocol: Restores corrupted or lost character execution contexts via standard replay hooks (char_state.py recover) without parsing the entire book text from scratch.
  • Cognitive Bias Injection: Equips specific character agents with localized sensory mapping errors and decision-making biases to eliminate homogenized AI dialogue and behavioral tropes.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*