fix(utils): classify runner URLs safely (#2191)

* fix(utils): classify runner URLs safely

* fix(utils): keep runner parse failures unknown

src/langbot/pkg/utils/runner.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+import ipaddress
+import re
 from urllib.parse import urlparse
 
 
@@ -44,6 +46,40 @@ LOCAL_PATTERNS = [
     '172.31.',
 ]
 
+HOST_LABEL_PATTERN = re.compile(r'^[a-z0-9](?:[a-z0-9-]{0,61}[a-z0-9])?$')
+
+
+def _is_valid_hostname(host: str) -> bool:
+    if host == 'localhost':
+        return True
+
+    try:
+        ipaddress.ip_address(host)
+        return True
+    except ValueError:
+        pass
+
+    if not host or len(host) > 253 or any(char.isspace() for char in host):
+        return False
+
+    host = host.rstrip('.')
+    if not host:
+        return False
+
+    return all(HOST_LABEL_PATTERN.match(label) for label in host.split('.'))
+
+
+def _is_local_host(host: str) -> bool:
+    if host == 'localhost':
+        return True
+
+    try:
+        ip_address = ipaddress.ip_address(host)
+    except ValueError:
+        return False
+
+    return ip_address.is_private or ip_address.is_loopback or ip_address.is_unspecified
+
 
 def get_runner_category(runner_name: str, runner_url: str) -> str:
     if not runner_url:
@@ -52,11 +88,14 @@ def get_runner_category(runner_name: str, runner_url: str) -> str:
     try:
         parsed_url = urlparse(runner_url)
         host = parsed_url.hostname.lower() if parsed_url.hostname else ''
+        _ = parsed_url.port
     except Exception:
         return RunnerCategory.UNKNOWN
 
-    for pattern in LOCAL_PATTERNS:
+    if not parsed_url.scheme or not host or not _is_valid_hostname(host):
-        if host.startswith(pattern):
+        return RunnerCategory.UNKNOWN
+
+    if _is_local_host(host):
         return RunnerCategory.LOCAL
 
     for domain in CLOUD_DOMAINS:

tests/unit_tests/utils/test_runner.py

@@ -0,0 +1,46 @@
+import pytest
+
+from langbot.pkg.utils.runner import RunnerCategory, get_runner_category
+
+
+@pytest.mark.parametrize(
+    'runner_url',
+    [
+        'api.dify.ai/v1',
+        'localhost:7860',
+        'https:///v1',
+        'https://',
+        'https://exa mple.com',
+        'http://[::1',
+        'http://localhost:bad',
+    ],
+)
+def test_get_runner_category_returns_unknown_for_invalid_urls(runner_url):
+    assert get_runner_category('dify-service-api', runner_url) == RunnerCategory.UNKNOWN
+
+
+@pytest.mark.parametrize(
+    'runner_url',
+    [
+        'http://localhost:7860',
+        'http://127.0.0.1:7860',
+        'http://10.0.0.1:7860',
+        'http://172.16.0.1:7860',
+        'http://172.31.255.255:7860',
+        'http://192.168.1.20:7860',
+        'http://[::1]:7860',
+    ],
+)
+def test_get_runner_category_detects_local_hosts_with_ipaddress(runner_url):
+    assert get_runner_category('langflow-api', runner_url) == RunnerCategory.LOCAL
+
+
+@pytest.mark.parametrize(
+    'runner_url',
+    [
+        'http://10.evil.com',
+        'http://192.168.example.com',
+    ],
+)
+def test_get_runner_category_does_not_treat_private_ip_prefix_domains_as_local(runner_url):
+    assert get_runner_category('langflow-api', runner_url) == RunnerCategory.CLOUD