Split canonical ReasoningEffort::Max out of Xhigh (providers P0c-1)

OpenAI (Responses) and Anthropic (Messages) treat xhigh and max as
DISTINCT effort levels in 2026, and the Kimi K3 wire's top tier is max —
the old parse alias (max→Xhigh) conflated them. Canonical Max now exists:
parse/as_str/serde split, Messages mapping sends xhigh and max as their
own tokens (was Xhigh→"max"), and the K3 menu token max carries
canonical Max end to end.

Kimi wire is byte-identical in all four flows (menu pick, restored
legacy xhigh session, --reasoning-effort flag, /effort command) —
adversarially traced and pinned: kimi_compat's string-level xhigh→max
rename covers legacy tokens, max passes through verbatim.

From the review:
- Rollback safety: persisted reasoning_effort (session summaries, chat
  history) deserializes leniently — unknown future tokens degrade to
  None with a warning instead of hiding sessions or failing resume.
- Restore migration: a pre-split xhigh override onto a model whose menu
  offers max but not xhigh (K3) migrates once, healing display/active-row
  drift and re-persisting the live vocabulary.
- /effort max now rejects (with the offered list) on models whose menu
  lacks a max row instead of silently applying xhigh; deliberate, tested.
- The interim Responses-backend Max→xhigh downgrade (async-openai has no
  Max variant through 0.41) warns loudly; real max wiring lands with the
  OpenAI provider cycle via post-serialize body patch.
- Two rusted ignored-e2e wire pins asserted the pre-adapt reasoning_effort
  key (deleted by the body adapter since ea0ce9d); they now pin the real
  thinking.effort=max shape.

crates/codegen/kigi-tui/src/acp/model_state.rs

@@ -247,9 +247,10 @@ impl ModelState {
         {
             return Some(option.value);
         }
-        // Canonical level (e.g. "high", "max"→xhigh) only if the model menu
-        // actually offers that value — not free-form power-user aliases that
-        // would 400 on the server (e.g. `none` on kigi-4.5).
+        // Canonical level (e.g. "high", "max") only if the model menu
+        // actually offers that value — not free-form power-user tokens that
+        // would 400 on the server (e.g. `none` on kigi-4.5, or `max` on a
+        // model whose vocabulary tops out at xhigh).
         let parsed = token.parse::<ReasoningEffort>().ok()?;
         options
             .iter()