[codex] add NCP semantic grounding

[jhull]

Summary

Adds NCP_SEMANTIC_GROUNDING.md and links it from the README for AI and tool consumers.

The new guide explains how NCP consumers should preserve Dramatica-informed meaning while reading, writing, or transforming protocol data. It covers:

• the boundary between the open interoperability layer and deeper Dramatica platform workflows
• the core distinction between Subtext, Storytelling, Story-level Moments, and Ideation
• common AI interpretation failures such as collapsing Main Character into Protagonist or rewriting Subtext from prose alone
• Holistic and character-framed affordance terminology, including Character Intentions, Character Repercussions, Dissonance, Relief, Friction, and Accommodation
• implementation guardrails for preserving canonical structural meaning while supporting writer-facing labels

Why

As more developers and writers feed NCP files into LLMs, the schema alone is not enough to prevent semantic drift. Models tend to flatten Dramatica terms into generic storytelling concepts: Main Character becomes hero, Objective Story becomes plot summary, Storyform becomes a beat sheet, and Holistic labels become vague emotional language.

This document gives adopters a public, protocol-safe grounding layer. It helps tools interoperate with NCP more accurately by explaining how to carry Dramatica-informed meaning across systems without turning the schema into a general-purpose theory manual or product workflow guide.

The goal is straightforward: make NCP easier for external tools and AI systems to consume while keeping terminology precise, portable, and aligned with the protocol.

Validation

• git diff --check HEAD~1..HEAD
• npm run validate:schema -> Schema validation checks passed (8 valid + 11 invalid fixtures).