Validate decoded base URL paths for unsafe segments

Co-authored-by: Shri Sukhani <shrisukhani@users.noreply.github.com>

Author: cursoragent

hyperbrowser/config.py

@@ -1,5 +1,5 @@
 from dataclasses import dataclass
-from urllib.parse import urlparse
+from urllib.parse import unquote, urlparse
 from typing import Dict, Mapping, Optional
 import os
 
@@ -58,6 +58,27 @@ def normalize_base_url(base_url: str) -> str:
             raise HyperbrowserError(
                 "base_url must not include query parameters or fragments"
             )
+
+        decoded_base_path = parsed_base_url.path
+        for _ in range(10):
+            next_decoded_base_path = unquote(decoded_base_path)
+            if next_decoded_base_path == decoded_base_path:
+                break
+            decoded_base_path = next_decoded_base_path
+        if "\\" in decoded_base_path:
+            raise HyperbrowserError("base_url must not contain backslashes")
+        if any(character.isspace() for character in decoded_base_path):
+            raise HyperbrowserError("base_url must not contain whitespace characters")
+        if any(
+            ord(character) < 32 or ord(character) == 127
+            for character in decoded_base_path
+        ):
+            raise HyperbrowserError("base_url must not contain control characters")
+        path_segments = [segment for segment in decoded_base_path.split("/") if segment]
+        if any(segment in {".", ".."} for segment in path_segments):
+            raise HyperbrowserError(
+                "base_url path must not contain relative path segments"
+            )
         return normalized_base_url
 
     @classmethod

tests/test_config.py

@@ -226,6 +226,10 @@ def test_client_config_rejects_empty_or_invalid_base_url():
         HyperbrowserError, match="base_url must not contain control characters"
     ):
         ClientConfig(api_key="test-key", base_url="https://example.local\x00api")
+    with pytest.raises(
+        HyperbrowserError, match="base_url path must not contain relative path segments"
+    ):
+        ClientConfig(api_key="test-key", base_url="https://example.local/%2e%2e/api")
 
 
 def test_client_config_normalizes_headers_to_internal_copy():
@@ -356,3 +360,15 @@ def test_client_config_normalize_base_url_validates_and_normalizes():
         HyperbrowserError, match="base_url must not contain control characters"
     ):
         ClientConfig.normalize_base_url("https://example.local\x00api")
+    with pytest.raises(
+        HyperbrowserError, match="base_url path must not contain relative path segments"
+    ):
+        ClientConfig.normalize_base_url("https://example.local/%252e%252e/api")
+    with pytest.raises(
+        HyperbrowserError, match="base_url must not contain backslashes"
+    ):
+        ClientConfig.normalize_base_url("https://example.local/%255Capi")
+    with pytest.raises(
+        HyperbrowserError, match="base_url must not contain whitespace characters"
+    ):
+        ClientConfig.normalize_base_url("https://example.local/%2520api")

tests/test_url_building.py

@@ -122,6 +122,13 @@ def test_client_build_url_rejects_runtime_invalid_base_url_changes():
         ):
             client._build_url("/session")
 
+        client.config.base_url = "https://example.local/%2e%2e/api"
+        with pytest.raises(
+            HyperbrowserError,
+            match="base_url path must not contain relative path segments",
+        ):
+            client._build_url("/session")
+
         client.config.base_url = "   "
         with pytest.raises(HyperbrowserError, match="base_url must not be empty"):
             client._build_url("/session")