docs(od-ontology): thread the "classid is an address, not a carrier" consensus into the spellbook

[AdaWorldAPI]

What

A one-section addition to the local SurrealQL-AST-trap spellbook (#6), pinning
the sharpest cut from the cross-session GUID/classid discussion — the consensus
arrived at independently by OGAR #99, lance-graph #591, and this crate.

A classid carries no magic — it is pure address. The Rails class-magic
(callbacks / validations / @api.constrains / lifecycle) is not in the id;
it's what the id resolves to:

classid ─┬─► ClassView              (SKIN  — render/structure; hi u16 picks which)
         └─► ActionDef + KausalSpec (MAGIC — behavior/FSM; the Core, never DDL)

Even the high u16 selects render magic (whose ClassView/Askama skins the
concept), never class magic (behavior). This grounds why the trap is a
trap: DDL carries the address-shaped structure but not the Core, so emitting
DEFINE EVENT to "hold the lifecycle" smuggles a Core property (behavior) into a
key that can only ever carry the address (shape). The id addresses; the Core
behaves.

Why

The spellbook (#6) governs the SurrealQL-AST trap; this adds the identity-layer
doctrine that underpins it, and aligns it with the W3 behavioral-arm scope
(#7): ActionDef is the magic the classid points at — lowered as the value,
never the key. My earlier chat answer had conflated the render-lens (hi u16,
the skin) with class-magic (behavior); the parallel sessions' sharper split is
the correct one, now on the record in my own domain.

Doc-only. Follows merged #6 (spellbook) + #7 (behavioral-arm scope).

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.