Roots vs Branches: Core Principles Persist Across Domains

Multi-domain creators (newsletters, client work, products, social) pay an 'identity tax' each time they start a new AI chat, reconstructing their thinking from scratch across separate Claude Projects. This fragments cognition: AI treats one mind as multiple personas, leading to inconsistent outputs that erode personal brand coherence. The fix distinguishes roots (stable psychological principles, philosophical defaults, aesthetic commitments true everywhere) from branches (tone, audience assumptions, pacing that adapt per context). Example: "Prioritize clarity over comprehensiveness" is a root manifesting as conversational LinkedIn posts, researched newsletters, or detailed specs. Rebuilding branches per project wastes time; inheriting roots once eliminates context-switching costs, backed by research showing task-switching reduces productivity (cited PMC study). Readers sense this inconsistency online when AI defaults to averages without your encoded principles.

Build a Root File in 20 Minutes for Instant Inheritance

Create a Markdown root file (300 words max) as the first layer in every Claude Project, skill, or agent. Paste it to load your universals: AI instantly knows how you reason, cutting clarifying questions and rephrasing. It saves calibration time, not tokens, by aligning AI to your defaults from the start. Distinctions:

• Vs. voice profile: Root captures how you think (decisions before style); voice handles how you write.
• Vs. context document: Root is prescriptive ("how to decide ambiguities"); context is descriptive (audience, goals).

To build: Pull writing from three different domains (e.g., newsletter + client email + product copy). Use the provided 4-phase prompt for extraction:

1. Pattern extraction (private): Spot recurring structures, commitments, aesthetics, reader outcomes.
2. Interview (6 targeted questions): Confirm deliberate patterns, costs of commitments, non-negotiables.
3. Pressure-test: Verify each principle appears everywhere (with maintenance cost) vs. adaptive branches.
4. Output structure:

   • Name's Root File

   • What this file is (inheritance explanation)

   • The roots (max 5: name + declarative principle + cost + success indicator)

   • What changes by context (branches)

   • How to use (inherit silently, flag drifts)

Declarative only, no hedging; fits one screen.

Author's Four Roots Power Consistent Outputs

Analyzing newsletter, notes, product copy, LinkedIn revealed these universals, now loaded in every project:

1. Strategy before execution: Diagnose thinking problems first; costs speed but yields better workflows.
2. Blueprints over fish: Deliver frameworks that generate context-specific answers; trades quick fixes for adaptability.
3. Intellectual respect as default: Assume reader smarts, explain machinery; narrows audience but builds loyalty.
4. Taste as non-negotiable filter: Applies uniform bar, adapts expression; refuses mediocrity despite platform pressures.

Result: One-time write, zero re-explanation. Monday switches (newsletter → client → roadmap) pay tax once upfront. Extends prior work like Dexter Protocol (modular files), Cleopatra Treaty (AI partnership), Crossword Method (central constraints). Download ready prompt from RobotsOS.