feat: add agent run queue controls

scripts/build-portal-runtime.mjs

@@ -346,6 +346,7 @@ async function writeMetadata() {
       '  node scripts/runtime-worker-drain.mjs undrain goosed-3',
       '  node scripts/check-tool-runtime.mjs',
       '  node scripts/check-agent-code-run-entry.mjs',
+      '  curl -sk https://mm.tkmind.cn/api/runtime/status  # includes toolRuntime.queue',
       '',
     ].join('\n'),
   );

scripts/runtime-slo-report.mjs

@@ -148,6 +148,7 @@ function summarizeRuntime(runtimeJson) {
     routerEnabled: Boolean(runtimeJson?.router?.enabled),
     publicBaseUrl: runtimeJson?.publicBaseUrl ?? null,
     toolRuntime: runtimeJson?.toolRuntime ?? null,
+    toolQueue: runtimeJson?.toolRuntime?.queue ?? null,
     workers: workers.map((worker) => ({
       id: worker.id,
       healthy: runtimeJson?.targets?.find((target) => target.target === worker.target)?.healthy ?? null,
@@ -220,6 +221,10 @@ for (const worker of runtimeSummary?.workers ?? []) {
   if (worker.firstTokenCount > 0 && worker.ewmaFirstTokenMs <= 0) failures.push(`${worker.id}_first_token_ewma_missing`);
 }
 if (runtimeSummary?.toolRuntime?.chatInjectsCodeTools !== false) failures.push('chat_injects_code_tools');
+if (runtimeSummary?.toolQueue?.error) failures.push('tool_queue_status_error');
+if (runtimeSummary?.toolQueue?.inFlight > runtimeSummary?.toolQueue?.maxConcurrentRuns) {
+  failures.push('tool_queue_concurrency_exceeded');
+}
 
 const report = {
   ok: failures.length === 0,