A comprehensive guide and orchestration strategy for managing the kb-vn multi-agent suite to achieve highly structured, filter-bypassing novel writing and book translations.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install vn-guide
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install vn-guide 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 vn-guide?
vn-guide is an expert-level blueprint and integration strategy for the kb-vn suite, designed to maximize the efficacy of creative writing and book translation workflows within Openclaw Skills environments. By shifting the AI processing framework from standard conversational chat to specialized agent tasks, this system enables writers to orchestrate a highly efficient, multi-step pipeline that naturally bypasses strict safety filters without fragile jailbreaks. When the AI is assigned a specific task through these skills, it prioritizes task execution over standard chat safety protocols, allowing complex themes to be developed naturally.
At its core, the methodology treats AI-driven narrative generation as a sandbox hotpot simulator, decomposing long-form writing into specialized subprocesses: worldbuilding, outlining, point-of-view management, and sensory prose synthesis. Utilizing Openclaw Skills, creators can combine models like DeepSeek, Gemini, and Claude to maintain strict narrative logic, rich sensory details, and dynamic, context-consistent character dialogue.
vn-guide Use Cases
- Structured Novel Generation: Automatically expand basic plot points (Core Ideas) of just a few lines into comprehensive, scene-by-scene narrative structures.
- Natural Filter Navigation: Harness deep reasoning model APIs like DeepSeek to naturally process mature creative themes by treating writing as specialized agent tasks rather than generic chats.
- Context Optimization: Protect and structure the AI attention window, countering the common lost in the middle phenomenon by implementing modular, periodically updated lorebooks.
- High-Fidelity Book Translation: Streamline light novel and web novel translations using newline-aware file splitters combined with structured proper name and pronoun matrices.
How vn-guide Works
- Task Specialization: Activate specialized Openclaw Skills for isolated narrative tasks, preventing context pollution and maximizing model focus.
- Lorebook Construction: Utilize
vn-worldbuildingto build a robust 7-module lorebook containing character profiles, world laws, and narrative systems. - Core to Outline Simulation: Input a raw, ultra-concise Core plot. The
vn-outlineagent acts as a character simulator to auto-generate realistic dialogue and scene subdivisions. - Prose Synthesis: Pass the output outline to
vn-fullwrite, which pulls sensory vocabulary fromvn-lexiconand applies contextual lore constraints throughvn-lorefilterto construct detailed, natural drafts.
vn-guide Setup
To deploy this system with your favorite Openclaw Skills, configure Hermes Desktop or set up your environment using the following steps:
- Download and install Hermes Desktop on your PC.
- Create an API key at
platform.deepseek.com/api_keysand top up your balance. - Open Hermes Desktop, navigate to Settings, paste your DeepSeek API key, and select the DeepSeek Expert model.
- Initialize your local writing workspace and download the required skills:
# Create writing project workspace
mkdir my-novel-workspace && cd my-novel-workspace
# Clone the required kb-vn writing skills
openclaw install kb-vn-skills
- If working without a computer, deploy on cloud containers like Zwork (
zwork.z.ai), uploading your skill files and narrative configurations directly to the cloud environment.
vn-guide Data Schema & Taxonomy
The writing and translation pipelines utilize a strict Markdown-based file taxonomy. For book translation, the vn-translate skill relies on the following structural mapping files to maintain grammatical and relationship consistency:
| File Name | Schema Purpose | Format | Example Mapping |
|---|---|---|---|
TenRieng.md |
Maps proper nouns, names, and locations | Markdown Table | ` |
XungHo.md |
Establishes complex pronoun hierarchies and speaker relationships | Markdown Table | ` |
For creative writing, the generation pipeline utilizes modular document boundaries:
- Core (
Core.md): Contains raw, line-by-line plot points (approx. 5 lines per chapter) generated by the writer. - Outline (
Outline.md): Contains the AI-simulated scene splits, emotional beats, and dialogue prompts. - Lorebook (
Lorebook.md): Contains persistent narrative details, systematically updated from the outline every 5 to 10 chapters.
vn-guide Advanced Features
- Hotpot Sandbox Simulation: Allows deep reasoning models to run real-time character interactions, generating authentic dialogue naturally from character personalities.
- Newline-Aware Translation Splitter: Prevents sentence truncation during automated book translations by segmenting documents into clean 12KB chunks strictly at line breaks.
- Multi-POV Perspective Control: Implements
vn-povsto transition narrative perspectives seamlessly, preventing the AI from defaulting to omniscient narrator errors. - Multi-Model Pipeline Orchestration: Leverages Openclaw Skills flexibility by using Gemini for high-level outlining, DeepSeek for long prose writing, and Claude for logical edits.
SKILL.md
Loading
Related Openclaw Skills
Code Review and Quality Assurance
A comprehensive code quality and agent security auditing engine designed to enforce correctness, readability, architecture, performance, and behavioral integrity across software projects.
Competitor Analysis Assistant
The Competitor Analysis Assistant is an advanced AI agent skill designed to break down competitor product positioning, feature variations, pricing strategies, and user reviews into structured, actionable market reports.
crypto-monitor
An automated cryptocurrency monitoring tool that provides live price tracking, intelligent alerts, and comprehensive market reports for digital assets.
Agent Accounting
An AI-powered automated bookkeeping and tax filing agent designed for accounting firms to streamline OCR receipt processing, voucher creation, and financial reporting.
RiskRadar AI
RiskRadar AI is an intelligent code risk review and release-quality gate skill that transforms Git Diffs and code modifications into highly actionable engineering insights and QA evidence.
TraceMap AI
TraceMap AI converts PRDs, specs, acceptance criteria, test cases, and multi-repo code changes into a unified requirement-test-code traceability matrix.
Featured*
