From 5f4a7e99f00de52b120485b4f2a34088bcfcb9a1 Mon Sep 17 00:00:00 2001
From: Teknium <127238744+teknium1@users.noreply.github.com>
Date: Sat, 8 Aug 2026 14:03:45 -0700
Subject: [PATCH] fix: explain provider DNS failures as possible offline state

---
 run_agent.py                                | 31 ++++++++++++++++--
 tests/run_agent/test_summarize_api_error.py | 35 ++++++++++++++++++++-
 2 files changed, 63 insertions(+), 3 deletions(-)

diff --git a/run_agent.py b/run_agent.py
index a8bbc515eaaa3..0aa009fae4b7c 100644
--- a/run_agent.py
+++ b/run_agent.py
@@ -2529,11 +2529,38 @@ class AIAgent:
         """Extract a human-readable one-liner from an API error.
 
         Handles Cloudflare HTML error pages (502, 503, etc.) by pulling the
-        <title> tag instead of dumping raw HTML.  Falls back to a truncated
-        str(error) for everything else.
+        <title> tag instead of dumping raw HTML. Network/DNS failures are
+        translated into an offline hint, including when an SDK wraps the
+        original OS error. Falls back to a truncated str(error) otherwise.
         """
         raw = str(error)
 
+        # Linux, macOS, and Windows use different low-level messages when DNS
+        # cannot resolve the provider while the device is offline. SDKs often
+        # wrap that OSError in a generic "Connection error", so inspect the
+        # exception chain before showing the top-level message to the user.
+        network_resolution_markers = (
+            "temporary failure in name resolution",
+            "name or service not known",
+            "nodename nor servname provided, or not known",
+            "getaddrinfo failed",
+            "no address associated with hostname",
+            "network is unreachable",
+        )
+        current: Optional[BaseException] = error
+        seen: set[int] = set()
+        while current is not None and id(current) not in seen:
+            seen.add(id(current))
+            if any(
+                marker in str(current).lower()
+                for marker in network_resolution_markers
+            ):
+                return (
+                    "Hermes can't reach the model provider. You may be offline. "
+                    "Check your internet connection and try again."
+                )
+            current = current.__cause__ or current.__context__
+
         if (
             isinstance(error, ValueError)
             and "expected ident at line" in raw.lower()
diff --git a/tests/run_agent/test_summarize_api_error.py b/tests/run_agent/test_summarize_api_error.py
index 82d9bcff1f7bd..d82c539a3b211 100644
--- a/tests/run_agent/test_summarize_api_error.py
+++ b/tests/run_agent/test_summarize_api_error.py
@@ -14,6 +14,7 @@ from types import SimpleNamespace
 from typing import Any
 
 import httpx
+import pytest
 
 from run_agent import AIAgent
 
@@ -37,9 +38,42 @@ def test_empty_body_falls_back_to_response_json_error_message():
     assert "HTTP 400" in summary
     assert "model `foo` does not exist" in summary
 
+@pytest.mark.parametrize(
+    "technical_message",
+    [
+        "Temporary failure in name resolution",
+        "Name or service not known",
+        "nodename nor servname provided, or not known",
+        "getaddrinfo failed",
+        "No address associated with hostname",
+        "Network is unreachable",
+    ],
+)
+def test_network_resolution_failure_explains_that_the_user_may_be_offline(
+    technical_message,
+):
+    error = OSError(-3, technical_message)
+
+    summary = AIAgent._summarize_api_error(error)
+
+    assert summary == (
+        "Hermes can't reach the model provider. You may be offline. "
+        "Check your internet connection and try again."
+    )
+    assert "name resolution" not in summary.lower()
 
 
+def test_wrapped_dns_resolution_failure_gets_the_same_friendly_message():
+    try:
+        try:
+            raise OSError(-3, "Temporary failure in name resolution")
+        except OSError as cause:
+            raise RuntimeError("Connection error.") from cause
+    except RuntimeError as error:
+        summary = AIAgent._summarize_api_error(error)
 
+    assert "You may be offline" in summary
+    assert "Connection error" not in summary
 
 
 def test_unread_streaming_response_does_not_crash_and_falls_back_to_exception_message():
@@ -62,4 +96,3 @@ def test_unread_streaming_response_does_not_crash_and_falls_back_to_exception_me
     summary = AIAgent._summarize_api_error(err)
     assert "HTTP 429" in summary
     assert "Gemini HTTP 429: quota exceeded" in summary
-