Subagents: a new chat reads /models again, so subagents switched on later show

The Subagents chip on a new chat follows `/models` -> model.subagents.offered, the same
team_run.offered the web page renders from. The app read /models once, at sign-in, and kept
it while it stayed open in the tray: subagents switched on afterwards (or a model added)
never reached a new chat until a restart. startNew() now re-reads it (at most every 5 s,
keeping the last list if the read fails) and redraws the draft's chips (modelsChanged()).
An open chat already re-reads its own `subagents` with GET /chats/{id}.

tests/ui: A5 cases -- parallel model offered: chip, slider to the server's cap, no note;
one-request model offered named_only: the `own` note; offered false everywhere: no chip;
switched on while open: the next new chat has it; an open chat follows its `subagents`.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

ui/js/chat.js

@@ -996,6 +996,18 @@ export function draftKind() {
   return draft?.kind || lastKind
 }
 
+/** `/models` was read again: a new chat being written is drawn with what it says now. */
+export function modelsChanged() {
+  if (!draft) return
+  const models = ctx.models()
+  if (!models.some((m) => m.id === draft.model)) {
+    const start = models.find((m) => m.default) || models[0]
+    draft.model = start?.id || ""
+    draft.effort = levelsOf(start).start
+  }
+  renderChips()
+}
+
 /** The greeting of a new chat, and the choice of what it is. */
 function renderHello() {
   const box = $("new-chat")