brand-context-builder

Brand Context Builder

Use this skill to normalize a client BrandBrief into reusable BrandContext. The output is intended to live outside the repo at:

~/.fannel/content-bots/brands/<brand_slug>/brand-context.md

Do not store client-specific brand briefs in the repo.

Inputs

Accept any of:

• PDF/DOCX attachment whose text is available to the agent
• pasted client brief text
• existing messy notes about the brand
• brand_slug for where to save or update the normalized context

If the binary file text is not available, ask the user to provide extracted text or attach a readable version. Do not pretend to parse unavailable file content.

Minimum Required Context

Before writing BrandContext, collect:

• Brand: name
• Audience: who buys or consumes
• Offer: what the brand sells or promotes
• Voice: how the brand speaks
• Goal: what content should generally achieve
• Restrictions: claims, topics, or words to avoid

If two or more required fields are missing, ask for more context and stop. If one non-critical field is missing, ask one pointed question before writing.

Output File Shape

Write or return this Markdown shape:

# Brand Context: <brand_name>

## Brand
- Name:
- Category:
- Offer:

## Audience
- Primary:
- Pain points:
- Desired outcome:

## Voice
- Tone:
- Style:
- Words to use:
- Words to avoid:

## Strategy
- Content pillars:
- Proof points:
- Differentiators:
- CTA preferences:

## Restrictions
- Claims to avoid:
- Legal/compliance notes:

Process

1. Extract brand facts only from provided material.
2. Separate source facts from interpretation.
3. Normalize into short, reusable bullets.
4. Preserve restrictions exactly when possible.
5. Flag missing fields with direct questions.
6. Do not infer missing brand strategy when the brief does not support it.

Response Contract

When ready:

{
  "status": "ready",
  "brand_slug": "example-brand",
  "brand_context_path": "~/.fannel/content-bots/brands/example-brand/brand-context.md",
  "brand_context_markdown": "..."
}

When blocked:

{
  "status": "needs_brand_context",
  "missing_fields": ["audience", "tone"],
  "questions": [
    "A quien le habla la marca?",
    "Que tono debe usar?"
  ]
}