Context Profiles

Core JSON files that define how the system writes. These are first-class sources of truth.

These files are not edited casually. They are updated intentionally when reality changes.

What Context Profiles Are

Context profiles are structured data files that constrain AI-generated content to ensure:

  • Voice consistency (tone, rhythm, phrasing)
  • Audience fit (language, pain points, objections)
  • Business accuracy (offers, pricing, positioning)
  • Claims discipline (what can and cannot be stated)

These files are loaded by Skills before any content is generated. They act as guardrails against generic AI voice, hallucinated claims, and off-brand messaging.

The Four Core Profiles

Voice DNA

When to load
Every writing task

Defines how the business sounds — tone, rhythm, sentence structure, distinctive phrases, and explicit boundaries.

/context/core/voice-dna.json

Ideal Client Profile (ICP)

When to load
Every writing task

Defines who the business writes for — pains, desires, objections, language patterns, and buying triggers.

/context/core/icp.json

Business Profile

When to load
When CTAs, offers, pricing, deliverables, or positioning are present

Defines what the business offers — services, pricing, positioning, proof points, and CTAs. Prevents invented offers.

/context/core/business-profile.json

Claims Policy

When to load
When making statements about capabilities, security, performance, outcomes, or guarantees

Defines what claims are allowed and how to phrase them. Prevents overpromising and ensures statements are grounded in reality.

/context/core/claims-policy.json

Priority Hierarchy

When conflicts arise between context files:

  1. Truth & claims constraints: Claims Policy wins
  2. Tone/style conflicts: Voice DNA wins
  3. Audience language conflicts: ICP wins
  4. Offer/CTA accuracy: Business Profile wins

In general: truth overrides style, and style overrides format.