A structured system to analyze authentic writing samples and generate a highly reproducible brand voice guide for automated workflows.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install voice-builder
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 voice-builder 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 Voice Builder?
The Voice Builder skill offers a systematic approach to extracting a brand's or individual's unique tone from real writing samples. Rather than relying on vague descriptors like 'engaging' or 'professional,' this system deconstructs text across multiple stylistic layers to output an actionable voice.md specification file. This ensures all downstream content tasks accurately mirror the original author's natural communication style.
By integrating this workflow, developers and creators can standardize voice parameters across various AI agent tasks. The output file integrates with other Openclaw Skills to maintain consistency, remove generic AI prose, and support scalable ghostwriting tasks.
Voice Builder Use Cases
- Deepening brand identity profiles once concrete writing samples become available.
- Solving issues where generated copy feels generic, robotic, or disconnected from the creator's true voice.
- Scaling ghostwriting tasks across multiple social channels while retaining authentic expression.
- Defining clear, objective guidelines for multi-agent workflows that require consistent brand messaging.
How Voice Builder Works
- Read Brand Profile: Analyze existing identity documents to align tone goals with audience expectations and core brand constraints.
- Gather Writing Evidence: Select 5 to 10 high-quality, authentic writing samples or spoken transcripts that represent the target voice.
- Deconstruct Style Layers: Evaluate the samples across lexicon, syntax, rhetoric, structure, stance, and negative space, citing exact quotes as evidence.
- Define the Fingerprint: Isolate 5 to 7 highly specific, reproducible signature traits alongside concrete rules on what style elements to avoid.
- Generate voice.md: Author the final specifications using standard structural templates, complete with a contextual tone map and practical do/don't rewrites.
- Validate and Refine: Run a validation test comparing generated test outputs against original samples, scoring the results to fine-tune the rule profile.
Voice Builder Setup
To configure and apply the voice-builder template, navigate to your workspace directory and ensure your reference guides are in place:
# Navigate to your agent workspace
cd my-agent-workspace
# Ensure the necessary reference directories exist
mkdir -p references
# Create or verify brand-profile.md exists before running
touch brand-profile.md
Ensure that you have access to companion Openclaw Skills to run validation pipelines and style checks seamlessly.
Voice Builder Data Schema & Taxonomy
The skill establishes a folder structure to maintain the voice specification workflow. Below is the primary taxonomy:
| Path | Purpose | Description |
|---|---|---|
brand-profile.md |
Prerequisite Input | Contains identity guidelines, target audience data, and broad guardrails. |
voice.md |
Core Deliverable | The generated rulebook containing the fingerprint, tone map, and do/don't rewrites. |
references/samples-guide.md |
Reference | Instructions on sourcing, vetting, and cleaning writing samples. |
references/analysis-framework.md |
Reference | The structural framework for analyzing style over six layers. |
references/voice-template.md |
Reference | The template structure used to generate the final voice guide. |
references/validation.md |
Reference | The score-based validation test process. |
Voice Builder Advanced Features
- Six-Layer Stylistic Analysis: Inspects lexical choice, structural patterns, syntax variance, rhetorical devices, authoritative stance, and negative space.
- Objective Verification Pipeline: Includes a step-by-step validation test to ensure the generated voice is empirically close to real samples.
- Contextual Tone Mapping: Defines explicit rules for how the voice dynamically shifts across different formats, platforms, and audiences.
- Robust Edge Case Handlers: Built-in processes for addressing thin sample size, voice drift, multilingual challenges, and aspirational tone transitions.
SKILL.md
Loading
Related Openclaw Skills
Brand Profile Skill
A foundational AI skill that builds a durable, reusable brand profile to ensure all downstream content generation matches a company's unique voice, audience, and guardrails.
Text Post and Microblog
A structured AI agent skill designed to craft high-impact, standalone text posts for modern microblogging platforms using the LINES framework.
Content Pillars
Content Pillars is an AI coding agent skill that analyzes brand strategy to generate 3 to 5 core ownable themes for content planning.
Hook Writer
An AI agent skill designed to analyze content, match brand voice, and write highly engaging, truthful opening lines optimized for specific platforms.
Science Research Writing
An AI agent skill designed to guide academic researchers through planning, drafting, revising, and auditing empirical research papers while preserving empirical evidence.
video-color-grade
A code-driven color grading tool that assesses footage, generates custom named looks, protects skin tones, and bakes portable 3D LUTs.
Featured*
