LLM Guidance
The system, put where the model can read it.
AI tools can produce YMF-brand-compliant output when given the right constraints. This section names the preferred tool, states the disclosure standard, and provides two prompts to copy verbatim — one for writing, one for visual design. Neither prompt is a suggestion; the model treats it as instruction.
Preferred model
Claude (Anthropic)
The prompts in this section were built and tested against Claude Sonnet (current). Claude reasons well about style constraints and holds persona instructions across long documents — both qualities matter for the YMF voice. Other models are not prohibited, but they require independent verification against these prompts before use in published material. Do not pin a specific model version; Anthropic updates Sonnet without changing the name, and newer versions consistently perform better.