fix(auto-reply/followup): surface billing notice on message_tool_only and reuse shared classifiers

Sanjays2402 · Sanjays2402 · commit 7e0c4adcd92e · 2026-07-02T10:57:56.000-07:00
Address review feedback on the followup billing/quota notice:

- Surface the notice on the default message_tool_only delivery path too.
  Previously the notice was suppressed when sourceReplyDeliveryMode was
  message_tool_only, which is exactly the path where a user would otherwise
  see dead air. Remove the suppression so a billing/quota/rate-limit failure
  always produces a short notice.

- Reuse the shared failover-matches classifiers (isBillingErrorMessage,
  isRateLimitErrorMessage, isPeriodicUsageLimitErrorMessage) instead of a
  narrower local substring list, so followup failure classification stays in
  lock-step with the rest of the runtime. The shared classifiers cover every
  case the local list did (billing, quota, insufficient_quota, rate limit,
  usage limit) plus more, and are more precise about weak signals.

Update the tests accordingly: the classifier test now uses a realistic
billing error string, and the message_tool_only test asserts the notice is
sent rather than suppressed.
diff --git a/src/auto-reply/reply/followup-runner.test.ts b/src/auto-reply/reply/followup-runner.test.ts
@@ -1695,7 +1695,7 @@ describe("createFollowupRunner quota/billing failure notice (#80700)", () => {
     expect(isFollowupQuotaBillingFailure("Third-party apps now draw from your extra usage")).toBe(
       true,
     );
-    expect(isFollowupQuotaBillingFailure("Provider anthropic has billing issue")).toBe(true);
+    expect(isFollowupQuotaBillingFailure("billing error: credit balance is too low")).toBe(true);
     expect(isFollowupQuotaBillingFailure("insufficient_quota: you exceeded your quota")).toBe(true);
     expect(isFollowupQuotaBillingFailure("429 rate limit reached")).toBe(true);
     expect(isFollowupQuotaBillingFailure("rate_limit_exceeded")).toBe(true);
@@ -1779,8 +1779,10 @@ describe("createFollowupRunner quota/billing failure notice (#80700)", () => {
     expect(onBlockReply).not.toHaveBeenCalled();
   });
 
-  it("suppresses the notice when sourceReplyDeliveryMode is message_tool_only", async () => {
-    runEmbeddedPiAgentMock.mockRejectedValueOnce(new Error("billing block from provider"));
+  it("sends the notice even when sourceReplyDeliveryMode is message_tool_only", async () => {
+    runEmbeddedPiAgentMock.mockRejectedValueOnce(
+      new Error("billing error: credit balance is too low"),
+    );
     const onBlockReply = vi.fn(async () => {});
     const runner = createFollowupRunner({
       opts: { onBlockReply },
@@ -1794,6 +1796,12 @@ describe("createFollowupRunner quota/billing failure notice (#80700)", () => {
 
     await runner(queued);
 
-    expect(onBlockReply).not.toHaveBeenCalled();
+    expect(onBlockReply).toHaveBeenCalledTimes(1);
+    const firstCall = (
+      onBlockReply.mock.calls as unknown as Array<Array<{ text?: string; isError?: boolean }>>
+    )[0];
+    const payload = requireRecord(firstCall?.[0], "notice payload");
+    expect(payload.text as string).toMatch(/billing\/quota\/rate-limit/);
+    expect(payload.isError).toBe(true);
   });
 });
diff --git a/src/auto-reply/reply/followup-runner.ts b/src/auto-reply/reply/followup-runner.ts
@@ -4,6 +4,11 @@ import { resolveBootstrapWarningSignaturesSeen } from "../../agents/bootstrap-bu
 import { resolveContextTokensForModel } from "../../agents/context.js";
 import { DEFAULT_CONTEXT_TOKENS } from "../../agents/defaults.js";
 import { runWithModelFallback } from "../../agents/model-fallback.js";
+import {
+  isBillingErrorMessage,
+  isPeriodicUsageLimitErrorMessage,
+  isRateLimitErrorMessage,
+} from "../../agents/pi-embedded-helpers/failover-matches.js";
 import { runEmbeddedPiAgent } from "../../agents/pi-embedded.js";
 import {
   buildAgentRuntimeDeliveryPlan,
@@ -40,21 +45,19 @@ type EmbeddedAgentRunResult = Awaited<ReturnType<typeof runEmbeddedPiAgent>>;
  * is a quota/billing/rate-limit class rejection. These are user-visible failure
  * modes where the originating channel should receive a short notice rather than
  * a silent drop. See https://github.com/openclaw/openclaw/issues/80700.
+ *
+ * Delegates to the shared failover-matches classifiers so this path stays in
+ * lock-step with the rest of the runtime's billing/rate-limit detection instead
+ * of maintaining a narrower local substring list.
  */
 export function isFollowupQuotaBillingFailure(message: string): boolean {
   if (!message) {
     return false;
   }
-  const lower = message.toLowerCase();
   return (
-    lower.includes("billing") ||
-    lower.includes("quota") ||
-    lower.includes("rate limit") ||
-    lower.includes("rate_limit") ||
-    lower.includes("rate-limit") ||
-    lower.includes("extra usage") ||
-    lower.includes("usage limit") ||
-    lower.includes("insufficient_quota")
+    isBillingErrorMessage(message) ||
+    isRateLimitErrorMessage(message) ||
+    isPeriodicUsageLimitErrorMessage(message)
   );
 }
 
@@ -382,10 +385,7 @@ export function createFollowupRunner(params: {
         const message = formatErrorMessage(err);
         replyOperation.fail("run_failed", err);
         defaultRuntime.error?.(`Followup agent failed before reply: ${message}`);
-        if (
-          isFollowupQuotaBillingFailure(message) &&
-          run.sourceReplyDeliveryMode !== "message_tool_only"
-        ) {
+        if (isFollowupQuotaBillingFailure(message)) {
           try {
             await sendFollowupPayloads(
               [
