What problem does it solve? Writing a complete long-form Vietnamese novel with AI requires orchestrating world-building, outlining, drafting, and editing across hundreds of chapters, which single-prompt approaches cannot sustain. This Skill automates that entire pipeline with three coordinated agents (Architect, Writer, Editor), rolling planning, and crash-safe checkpointing. ## Core Features & Use Cases - Autonomous Multi-Agent Writing: Coordinates Architect, Writer, and Editor agents through one LLM loop to produce 500+ chapter novels from a single prompt, with 3-tier context management and 7-dimension quality evaluation. - Checkpoint Recovery & Intervention: Auto-saves after every step so crashes, network loss, or Ctrl+C never lose progress, and lets you inject real-time feedback without stopping the process. - Multi-Provider LLM Support: Works with OpenRouter, Anthropic, Gemini, OpenAI, Deepseek, and local Ollama models, with per-role model assignment to balance cost and quality. - Use Case: A writer wants an 80-chapter Vietnamese palace-intrigue romance novel. They launch the TUI with a one-sentence prompt, add a custom_rules.md defining style and character constraints, let the system write autonomously, intervene when pacing drifts, and export the finished work to EPUB. ## Quick Start Install ainovel-cli via Docker, configure an LLM provider in config.json, then launch the TUI and enter a Vietnamese story prompt to begin autonomous novel generation.