From c6db48b346f919d045e3d97ace7ee810379eadc7 Mon Sep 17 00:00:00 2001
From: Shadow <hi@shadowing.dev>
Date: Mon, 26 Jan 2026 19:47:18 -0600
Subject: [PATCH] fix: harden unhandled rejection handling and tts menus
 (#2451) (thanks @Glucksberg)

---
 CHANGELOG.md                                 |  2 ++
 src/auto-reply/reply/commands-tts.ts         |  6 +++---
 src/auto-reply/reply/commands.test.ts        | 14 ++++++++++++++
 src/auto-reply/reply/dispatch-from-config.ts | 10 ++++++++--
 4 files changed, 27 insertions(+), 5 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index ed99095aa..8abb35bbd 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -55,6 +55,8 @@ Status: unreleased.
 - **BREAKING:** Gateway auth mode "none" is removed; gateway now requires token/password (Tailscale Serve identity still allowed).
 
 ### Fixes
+- Gateway: suppress AbortError and transient network errors in unhandled rejections. (#2451) Thanks @Glucksberg.
+- TTS: keep /tts status replies on text-only commands and avoid duplicate block-stream audio. (#2451) Thanks @Glucksberg.
 - Security: pin npm overrides to keep tar@7.5.4 for install toolchains.
 - Security: properly test Windows ACL audit for config includes. (#2403) Thanks @dominicnunez.
 - CLI: recognize versioned Node executables when parsing argv. (#2490) Thanks @David-Marsh-Photo.
diff --git a/src/auto-reply/reply/commands-tts.ts b/src/auto-reply/reply/commands-tts.ts
index bba7e2b02..04b60a4e9 100644
--- a/src/auto-reply/reply/commands-tts.ts
+++ b/src/auto-reply/reply/commands-tts.ts
@@ -25,11 +25,11 @@ type ParsedTtsCommand = {
 };
 
 function parseTtsCommand(normalized: string): ParsedTtsCommand | null {
-  // Accept `/tts <action> [args]` - return null for `/tts` alone to trigger inline menu.
-  if (normalized === "/tts") return null;
+  // Accept `/tts` and `/tts <action> [args]` as a single control surface.
+  if (normalized === "/tts") return { action: "status", args: "" };
   if (!normalized.startsWith("/tts ")) return null;
   const rest = normalized.slice(5).trim();
-  if (!rest) return null;
+  if (!rest) return { action: "status", args: "" };
   const [action, ...tail] = rest.split(/\s+/);
   return { action: action.toLowerCase(), args: tail.join(" ").trim() };
 }
diff --git a/src/auto-reply/reply/commands.test.ts b/src/auto-reply/reply/commands.test.ts
index 7078c15dc..fd8236c95 100644
--- a/src/auto-reply/reply/commands.test.ts
+++ b/src/auto-reply/reply/commands.test.ts
@@ -420,3 +420,17 @@ describe("handleCommands subagents", () => {
     expect(result.reply?.text).toContain("Status: done");
   });
 });
+
+describe("handleCommands /tts", () => {
+  it("returns status for bare /tts on text command surfaces", async () => {
+    const cfg = {
+      commands: { text: true },
+      channels: { whatsapp: { allowFrom: ["*"] } },
+      messages: { tts: { prefsPath: path.join(testWorkspaceDir, "tts.json") } },
+    } as ClawdbotConfig;
+    const params = buildParams("/tts", cfg);
+    const result = await handleCommands(params);
+    expect(result.shouldContinue).toBe(false);
+    expect(result.reply?.text).toContain("TTS status");
+  });
+});
diff --git a/src/auto-reply/reply/dispatch-from-config.ts b/src/auto-reply/reply/dispatch-from-config.ts
index 146e39f57..1dcd770bc 100644
--- a/src/auto-reply/reply/dispatch-from-config.ts
+++ b/src/auto-reply/reply/dispatch-from-config.ts
@@ -16,7 +16,7 @@ import { formatAbortReplyText, tryFastAbortFromMessage } from "./abort.js";
 import { shouldSkipDuplicateInbound } from "./inbound-dedupe.js";
 import type { ReplyDispatcher, ReplyDispatchKind } from "./reply-dispatcher.js";
 import { isRoutableChannel, routeReply } from "./route-reply.js";
-import { maybeApplyTtsToPayload, normalizeTtsAutoMode } from "../../tts/tts.js";
+import { maybeApplyTtsToPayload, normalizeTtsAutoMode, resolveTtsConfig } from "../../tts/tts.js";
 
 const AUDIO_PLACEHOLDER_RE = /^<media:audio>(\s*\([^)]*\))?$/i;
 const AUDIO_HEADER_RE = /^\[Audio\b/i;
@@ -342,10 +342,16 @@ export async function dispatchReplyFromConfig(params: {
       }
     }
 
+    const ttsMode = resolveTtsConfig(cfg).mode ?? "final";
     // Generate TTS-only reply after block streaming completes (when there's no final reply).
     // This handles the case where block streaming succeeds and drops final payloads,
     // but we still want TTS audio to be generated from the accumulated block content.
-    if (replies.length === 0 && blockCount > 0 && accumulatedBlockText.trim()) {
+    if (
+      ttsMode === "final" &&
+      replies.length === 0 &&
+      blockCount > 0 &&
+      accumulatedBlockText.trim()
+    ) {
       try {
         const ttsSyntheticReply = await maybeApplyTtsToPayload({
           payload: { text: accumulatedBlockText },