fix(gateway): AUTO 范围冲突的提示要放在 checkQuota，否则是死代码

上一版把消息加在请求路径的范围检查上，实测才发现它永远走不到：
AUTO 请求先经过 checkQuota(ctx, "AUTO")（在链被读取之前），范围不含 AUTO
时它当场返回旧文案「model \"AUTO\" is not allowed for this key」——那句会
把运维引去查配额，而配额并不是原因。

改为在 checkQuota 内对 AUTO 分流，输出说明真实原因与修法。请求路径那处
换成注释说明为何不再重复。

判据 1 条，断言 checkQuota 函数体内必须同时出现 isAuto(model) 与新文案
——放在别处不算数。变异（删掉该分支）导致编译失败，被捕获。

生产实测（临时把某密钥范围改成不含 AUTO 再恢复）：
  HTTP 403  model_not_allowed
  model \"AUTO\" is not in this key's model scope, so it cannot use AUTO;
  add \"AUTO\" to the key's models or remove the scope restriction

Co-Authored-By: ModelRouter <noreply@modelrouter.dev>

internal/gateway/chat.go

@@ -256,7 +256,17 @@ func (g *Gateway) checkQuota(ctx context.Context, model string) *quotaRejection
 		}
 		return nil
 	}
-	return &quotaRejection{msg: fmt.Sprintf("model %q is not allowed for this key", model)}
+// A scope that simply does not list the model is a different problem from
+// a quota being exhausted, and the message has to say which: telling an
+// operator "not allowed for this key" when they configured a quota sends
+// them to the wrong setting. AUTO is checked here before the chain is even
+// consulted, so this is the message that surfaces for AUTO.
+if isAuto(model) {
+return &quotaRejection{msg: fmt.Sprintf(
+"model %q is not in this key's model scope, so it cannot use AUTO; "+
+"add %q to the key's models or remove the scope restriction", model, model)}
+}
+return &quotaRejection{msg: fmt.Sprintf("model %q is not allowed for this key", model)}
 }
 
 // quotaWindowSuffix describes a quota's reset window for an error message, so
@@ -452,19 +462,11 @@ func (g *Gateway) handleChat(w http.ResponseWriter, r *http.Request) {
 			return
 		}
 	}
-	// A key whose model scope does not include AUTO cannot use AUTO at all,
-	// even when it has its own AUTO chain configured. That combination is
-	// easy to set up by accident and produced a misleading error: the request
-	// fell through to the generic "model %q is not configured" below, which
-	// claims AUTO does not exist — it does, this key just may not use it. Name
-	// the actual reason so the admin can fix the scope.
-	if isAuto(model) {
-		if allow := g.allowedModels(r.Context()); allow != nil && !g.hasScopeModel(allow, model) {
-			writeError(w, http.StatusForbidden, "model_not_allowed",
-				fmt.Sprintf("model %q is not in this key's model scope, so it cannot use AUTO; add %q to the key's models or remove the scope restriction", model, model))
-			return
-		}
-	}
+	// Note: a key whose scope excludes AUTO is rejected earlier, by
+	// checkQuota(ctx, "AUTO") above, which produces the same message. That
+	// ordering matters — checkQuota runs before the chain is consulted, so it
+	// is the check an AUTO request actually hits. Duplicating it here would be
+	// dead code.
 	cands, effective := g.resolveCands(r.Context(), &req)
 	if len(cands) == 0 {
 		writeError(w, http.StatusNotFound, "model_not_found", fmt.Sprintf("model %q is not configured", model))