cf-style-guide

Extracts brand voice, terminology, and compliance guardrails from style guides into a structured brand profile JSON.

1|Updated Aug 18, 2026
One-click install
npx skills add https://github.com/teachskillofskills-ai/ContentForge-techshu --skill cf-style-guide-teachskillofskills-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cf-style-guide
Source: https://github.com/teachskillofskills-ai/ContentForge-techshu/tree/main/skills/cf-style-guide
Command: npx skills add https://github.com/teachskillofskills-ai/ContentForge-techshu --skill cf-style-guide-teachskillofskills-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Marketing teams onboarding a new brand must manually translate style guide documents into machine-usable rules for AI content pipelines, risking inconsistent voice, missed banned terms, and compliance violations. This Skill converts an existing style guide — a .docx/.pdf document, a URL, or manual input — into a structured brand profile JSON that every ContentForge pipeline phase reads. ## Core Features & Use Cases - Voice and Terminology Extraction: Parses tone, formality level (1-5), personality traits, approved/banned/conditional terms, and acronyms from documents, URLs (via WebFetch), or interactive manual input. - Compliance Guardrail Parsing: Extracts required disclaimers, prohibited claims, regulatory rules (HIPAA, FDA, FTC), and sensitivity guidelines for regulated industries like healthcare and finance. - Tracking Backend Configuration: Configures Google Sheets + Drive, Airtable, or local filesystem tracking, with environment probes for Cowork sandbox persistence via Drive sync. - Use Case: A healthcare marketing team imports their 40-page brand guidelines PDF; the Skill extracts 47 approved terms, 23 banned terms, and 4 required disclaimers, validates pipeline compatibility, and saves the profile for use with /contentforge:create-content --brand=AcmeMed. ## Quick Start Ask the AI to import your brand guidelines by running /contentforge:cf-style-guide with your brand name and a style guide URL or document path.

Frequently Asked Questions about cf-style-guide

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I import a brand style guide into an AI content pipeline?

Run /contentforge:cf-style-guide with a brand name and a source — a URL, a .docx/.pdf file, or manual input. The Skill extracts voice, terminology, and guardrails, then saves a validated brand profile JSON that pipeline phases read for drafting and compliance checks.

What file formats can a style guide importer parse?

This Skill parses .docx and .pdf documents, fetches web pages and Notion-style URLs via WebFetch, and accepts fully manual interactive input. PDF parsing can miss complex multi-column layouts, so converting to .docx first gives better results.

Can I import only terminology or only compliance guardrails?

Yes, use the --scope flag with values voice, terminology, guardrails, or all (the default). For example, --scope=guardrails extracts only disclaimers, prohibited claims, and regulatory rules from a compliance document.

Does the brand profile work with Google Drive and Airtable?

Yes, the tracking backend step supports Google Sheets + Drive via service account or MCP, Airtable via a personal access token, or a local filesystem default. In Cowork sandbox environments, Drive sync persists profiles across sessions.

Why does style guide extraction return low confidence or zero terms?

Low confidence occurs when the guide lacks explicit voice sections or embeds terminology in prose rather than lists. Supplement with --source=manual to provide voice characteristics interactively, or use --scope=terminology with manual input for missing terms.

What are the limitations of automated style guide import?

Documents are limited to 50 pages or 25,000 words, non-English guides have less accurate terminology extraction, and visual identity sections like logos and fonts are skipped. Implicit voice shown only through examples also lowers extraction confidence.