feat(display): add suppress_retry_status platform config flag

gateway/display_config.py

@@ -41,6 +41,12 @@
     # live, just cleaned up after success so the chat doesn't fill up with
     # stale breadcrumbs. Failed runs leave bubbles in place as breadcrumbs.
     "cleanup_progress": False,
+    # When true, suppresses retry/empty-response/thinking-only/fallback status
+    # bubbles and the "(empty)" sentinel substitution. Intended for deployments
+    # where agent personas can legitimately produce empty responses (e.g. "stay
+    # silent when addressed to someone else") that the generic retry logic would
+    # otherwise misinterpret as failures. Off by default.
+    "suppress_retry_status": False,
 }
 
 # ---------------------------------------------------------------------------
@@ -194,7 +200,7 @@ def _normalise(setting: str, value: Any) -> Any:
         if isinstance(value, str):
             return value.lower() in {"true", "1", "yes", "on"}
         return bool(value)
-    if setting == "cleanup_progress":
+    if setting in {"cleanup_progress", "suppress_retry_status"}:
         if isinstance(value, str):
             return value.lower() in {"true", "1", "yes", "on"}
         return bool(value)

gateway/run.py

@@ -1073,11 +1073,33 @@ def _format_gateway_process_notification(evt: dict) -> "str | None":
 _gateway_runner_ref: _weakref.ref = lambda: None
 
 
+def _resolve_suppress_retry_status(platform_key: str) -> bool:
+    """Resolve the ``display.platforms.<platform>.suppress_retry_status`` flag.
+
+    Returns True when the active platform has ``suppress_retry_status`` set to a
+    truthy value in the display config.  Used to silence the ⚠️ empty-response
+    substitution and the generic "no response was generated" warning for
+    deployments where empty model responses are a legitimate outcome (e.g. an
+    agent persona that stays silent when a message is addressed to another
+    participant).
+    """
+    try:
+        from gateway.display_config import resolve_display_setting
+        from hermes_cli.config import load_config as _load_cfg
+        cfg = _load_cfg() or {}
+        return bool(
+            resolve_display_setting(cfg, platform_key, "suppress_retry_status", False)
+        )
+    except Exception:
+        return False
+
+
 def _normalize_empty_agent_response(
     agent_result: dict,
     response: str,
     *,
     history_len: int = 0,
+    platform_key: str = "",
 ) -> str:
     """Normalize empty/None agent responses into user-facing messages.
 
@@ -1111,6 +1133,8 @@ def _normalize_empty_agent_response(
         if agent_result.get("partial"):
             err = agent_result.get("error", "processing incomplete")
             return f"⚠️ Processing stopped: {str(err)[:200]}. Try again."
+        if _resolve_suppress_retry_status(platform_key):
+            return response
         return (
             "⚠️ Processing completed but no response was generated. "
             "This may be a transient error — try sending your message again."
@@ -7644,11 +7668,14 @@ async def _handle_message_with_agent(self, event, source, _quick_key: str, run_g
             # prefill, empty-retry, fallback).  Sending the raw sentinel
             # looks like a bug; a short explanation is more helpful.
             if response == "(empty)":
-                response = (
-                    "⚠️ The model returned no response after processing tool "
-                    "results. This can happen with some models — try again or "
-                    "rephrase your question."
-                )
+                if _resolve_suppress_retry_status(_platform_name):
+                    response = ""
+                else:
+                    response = (
+                        "⚠️ The model returned no response after processing tool "
+                        "results. This can happen with some models — try again or "
+                        "rephrase your question."
+                    )
             agent_messages = agent_result.get("messages", [])
             _response_time = time.time() - _msg_start_time
             _api_calls = agent_result.get("api_calls", 0)
@@ -7681,6 +7708,7 @@ async def _handle_message_with_agent(self, event, source, _quick_key: str, run_g
             # the case where agent did work but returned no text. Fix for #18765.
             response = _normalize_empty_agent_response(
                 agent_result, response, history_len=len(history),
+                platform_key=_platform_name,
             )
 
             # If the agent's session_id changed during compression, update

run_agent.py

@@ -2884,6 +2884,38 @@ def _should_emit_quiet_tool_messages(self) -> bool:
             and getattr(self, "platform", "") == "cli"
         )
 
+    def _should_suppress_retry_status(self) -> bool:
+        """Return True when retry/empty-response status bubbles should be suppressed.
+
+        Reads ``display.platforms.<platform>.suppress_retry_status`` from config.
+        Cached per agent instance.
+        """
+        cached = getattr(self, "_suppress_retry_status_cached", None)
+        if cached is not None:
+            return cached
+        try:
+            from gateway.display_config import resolve_display_setting
+            from hermes_cli.config import load_config as _load_cfg
+            cfg = _load_cfg() or {}
+            platform_key = (self.platform or "").lower().strip() or "cli"
+            result = bool(
+                resolve_display_setting(cfg, platform_key, "suppress_retry_status", False)
+            )
+        except Exception:
+            result = False
+        self._suppress_retry_status_cached = result
+        return result
+
+    def _emit_retry_status(self, message: str) -> None:
+        """Emit a retry/empty-response/fallback status bubble, if not suppressed.
+
+        No-op when ``suppress_retry_status`` is enabled for the active platform.
+        Otherwise identical to ``_emit_status``.
+        """
+        if self._should_suppress_retry_status():
+            return
+        self._emit_status(message)
+
     def _emit_status(self, message: str) -> None:
         """Emit a lifecycle status message to both CLI and gateway channels.
 
@@ -15091,7 +15123,7 @@ def _stop_spinner():
                                 "Empty response after tool calls — nudging model "
                                 "to continue processing"
                             )
-                            self._emit_status(
+                            self._emit_retry_status(
                                 "⚠️ Model returned empty after tool calls — "
                                 "nudging to continue"
                             )
@@ -15137,7 +15169,7 @@ def _stop_spinner():
                                 "prefilling to continue (%d/2)",
                                 self._thinking_prefill_retries,
                             )
-                            self._emit_status(
+                            self._emit_retry_status(
                                 f"↻ Thinking-only response — prefilling to continue "
                                 f"({self._thinking_prefill_retries}/2)"
                             )
@@ -15173,7 +15205,7 @@ def _stop_spinner():
                                 "retry %d/3 (model=%s)",
                                 self._empty_content_retries, self.model,
                             )
-                            self._emit_status(
+                            self._emit_retry_status(
                                 f"⚠️ Empty response from model — retrying "
                                 f"({self._empty_content_retries}/3)"
                             )
@@ -15192,13 +15224,13 @@ def _stop_spinner():
                                 self._empty_content_retries, self.model,
                                 self.provider,
                             )
-                            self._emit_status(
+                            self._emit_retry_status(
                                 "⚠️ Model returning empty responses — "
                                 "switching to fallback provider..."
                             )
                             if self._try_activate_fallback():
                                 self._empty_content_retries = 0
-                                self._emit_status(
+                                self._emit_retry_status(
                                     f"↻ Switched to fallback: {self.model} "
                                     f"({self.provider})"
                                 )
@@ -15232,7 +15264,7 @@ def _stop_spinner():
                                 "after exhausting retries and fallback. "
                                 "Reasoning: %s", reasoning_preview,
                             )
-                            self._emit_status(
+                            self._emit_retry_status(
                                 "⚠️ Model produced reasoning but no visible "
                                 "response after all retries. Returning empty."
                             )
@@ -15244,7 +15276,7 @@ def _stop_spinner():
                                 self._empty_content_retries, self.model,
                                 self.provider,
                             )
-                            self._emit_status(
+                            self._emit_retry_status(
                                 "❌ Model returned no content after all retries"
                                 + (" and fallback attempts." if self._fallback_chain else
                                    ". No fallback providers configured.")