story-pipeline

What This Skill Does

The story-pipeline skill is a robust framework for long-form narrative generation within the OpenClaw ecosystem. Designed for writers, content creators, and story architects, it enables the creation of multi-episode stories by maintaining long-term memory via a sophisticated graph management system. Unlike simple text-generation prompts, this skill enforces a structured 'Pipeline' approach where every episode is subjected to a dual-gate validation mechanism: an internal AI quality review and an essential human-in-the-loop confirmation phase. By automatically tracking characters, locations, and narrative hooks through a remote graph API, it ensures continuity across up to ten or more episodes. It effectively handles complex character growth, emotional pacing, and plot device management, making it an ideal tool for writing novels, screenplays, or serial fiction.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/hexidyg/story-master Ensure that your environment is configured for external network access as this skill communicates with a graph database endpoint for state persistence.

Use Cases

• Long-form Narrative Writing: Generate structured, multi-part serialized stories with consistent internal logic.
• Interactive Roleplay Systems: Maintain world states for persistent tabletop-style roleplaying scenarios.
• Content Creation Workflows: Streamline the production of web-series or blog-serialized fiction where consistency is paramount.
• Creative Brainstorming: Use the automated hook management to explore potential plot directions for an existing core theme.

Example Prompts

1. "开始一个新的剧情管道，主题：中国小女孩逆袭，目标集数：10集"
2. "继续剧情管道 pipeline_2026-03-05-001"
3. "通过，继续下一集"

Tips & Limitations

• Graph Persistence: The skill relies on an external API endpoint (framedream.art). Ensure a stable internet connection for graph queries and storage.
• AI Quality Gate: The system automatically triggers a retry if the AI quality score falls below 7. You can influence the generation by providing detailed feedback during the 'Need modification' phase.
• State Management: Keep an eye on data/pipeline_state.json. You can manually check this file to verify the progress and current status of your active pipelines.
• Limitations: The skill is designed for serialized narratives. Extremely complex, non-linear world-building with thousands of entities may require manual oversight to ensure the graph database captures every nuance.