Brand Context

Build and maintain one brand-context.md file per brand: the single source of truth for what the business sells, who it sells to, how it talks, and what it can prove. Every other skill in this collection reads this file before asking the user brand questions, so the user explains their business once — not once per task.

This is the hub in a hub-and-spoke model. Ad copy, blog posts, cold email, social content, and landing-page work are all downstream of the same positioning and voice; when those live in one maintained doc, every downstream skill starts warm and stays consistent.

The file contract. The doc lives in persistent storage at a stable path — /files/<brand-slug>/brand-context.md — and is reused at that SAME path every time. Create it once, then update it in place; never recreate it from scratch when it already exists (you would destroy verbatim customer language and proof points the user curated). If a run touches the doc, bump the Last updated line.

Requirements

• Hyper MCP connected. https://app.hyperfx.ai/mcp — needed for the file tools that persist the doc, and for auto-draft.
• Recommended for auto-draft, enabled at https://app.hyperfx.ai/apps: Firecrawl (read the brand's site and extract visual identity). Without it, skip auto-draft and use the interview path — the skill still works.
• Optional enrichment: the Apify scrapers used by customer-research (verbatim customer language) and HyperSEO (competitive landscape). Both are nice-to-have; mark any section built without real data as unverified rather than inventing content.

Tool surface