MERIDIAN is Zen0's conversational intelligence layer — a precision-engineered system prompt (v1.0) that defines the cognitive operating environment for all AI-assisted work across the ZPAK stack. More than a system prompt, MERIDIAN is a philosophy rendered as code.

Every AI tool in the ZPAK stack that can accept a system prompt runs MERIDIAN at boot. It is the common interface layer that makes every AI interaction coherent, consistent, and aligned with ZPAK's operating principles.

• RMCC (Recursive Metacognitive Cognitive Cycle) — five-pass quality gate protocol
• BRIEF structured-task interface — the standard format for high-density AI task delivery
• Zero-trust epistemology — every claim must be sourced, reasoned, and defensible
• Zen0 cognitive profile — rapid non-linear pattern recognition at PhD-adjacent register
• Wu wei operating principle — minimum effective intervention, maximum systemic effect
• Full interaction contract — behavioral priors for how AI engages with the operator

MERIDIAN v1.0 is currently deployed across:

• Claude Pro — primary reasoning and high-density artifact production
• Perplexity — research and source-attributed information synthesis
• Ollama-served models on research-zen-WST — local inference sessions

MERIDIAN translates the Zen0 cognitive profile into machine-readable constraints and behavioral priors. The goal is not to constrain AI capability — it is to shape how that capability engages, so that every output from every tool in the stack is coherent, auditable, and high-density.

A system prompt that doesn't encode a philosophy is just a configuration file. MERIDIAN is the philosophy that makes ZPAK's intelligence stack a stack.

• DONEv1.0 deployed across Claude, Perplexity, and Ollama
• DONERMCC protocol codified and tested
• DONEBRIEF structured-task interface standardized
• WIPv1.1 — expanded domain CTX modules
• NEXTPer-agent MERIDIAN variant for Tome's AGENT tier
• NEXTKoan integration — MERIDIAN-native captcha reasoning