diff --git a/benchmarks/runner.py b/benchmarks/runner.py
index 5b837abb..97769cd4 100644
--- a/benchmarks/runner.py
+++ b/benchmarks/runner.py
@@ -263,9 +263,7 @@ def _update_progress(
 
         # Call custom callback if provided
         if self.config.progress_callback:
-            self.config.progress_callback(
-                stats.completed, stats.total, sample_result, stats
-            )
+            self.config.progress_callback(stats.completed, stats.total, sample_result, stats)
 
         # Update display based on mode
         if progress_mode == "tqdm" and pbar is not None:
@@ -282,7 +280,7 @@ def _update_progress(
             if stats.completed % interval == 0 or stats.completed == stats.total:
                 print(
                     f"  Progress: {stats.completed}/{stats.total} "
-                    f"({stats.completed/stats.total:.0%}) | "
+                    f"({stats.completed / stats.total:.0%}) | "
                     f"Acc: {stats.accuracy:.1%} | "
                     f"Errors: {stats.errors} | "
                     f"ETA: {self._format_eta(stats.eta_seconds)}"
diff --git a/examples/oolong_example.py b/examples/oolong_example.py
index 69f3043b..b5ca0634 100644
--- a/examples/oolong_example.py
+++ b/examples/oolong_example.py
@@ -32,9 +32,7 @@ def load_random_oolong_row() -> dict:
 
 def main():
     # Parse command-line arguments
-    parser = argparse.ArgumentParser(
-        description="Run Oolong benchmark example with RLM"
-    )
+    parser = argparse.ArgumentParser(description="Run Oolong benchmark example with RLM")
     parser.add_argument(
         "--backend",
         type=str,
diff --git a/examples/subprocess_repl_demo.py b/examples/subprocess_repl_demo.py
index 17031c65..2642edfd 100644
--- a/examples/subprocess_repl_demo.py
+++ b/examples/subprocess_repl_demo.py
@@ -47,6 +47,7 @@ def demonstrate_isolation():
     print(f"Platform: {platform.system()}")
     if platform.system() == "Linux":
         import shutil
+
         if not shutil.which("bwrap"):
             print("Note: bubblewrap not installed - filesystem sandbox disabled")
     print()
diff --git a/rlm/environments/subprocess_repl.py b/rlm/environments/subprocess_repl.py
index 0ceef81f..bb6c3106 100644
--- a/rlm/environments/subprocess_repl.py
+++ b/rlm/environments/subprocess_repl.py
@@ -298,9 +298,22 @@ def __init__(
 
         # Pre-approved packages (stdlib + user-specified)
         self.allowed_packages: set[str] = {
-            "json", "re", "math", "collections", "itertools",
-            "functools", "datetime", "random", "string", "typing",
-            "os", "sys", "io", "time", "pathlib", "copy",
+            "json",
+            "re",
+            "math",
+            "collections",
+            "itertools",
+            "functools",
+            "datetime",
+            "random",
+            "string",
+            "typing",
+            "os",
+            "sys",
+            "io",
+            "time",
+            "pathlib",
+            "copy",
         }
         if allowed_packages:
             self.allowed_packages.update(allowed_packages)
@@ -472,10 +485,12 @@ def _install_package(self, package: str):
             capture_output=True,
         )
         self._installed_packages.add(package)
-        self._overhead_stats["package_installs"].append({
-            "package": package,
-            "time_ms": (time.perf_counter() - start) * 1000,
-        })
+        self._overhead_stats["package_installs"].append(
+            {
+                "package": package,
+                "time_ms": (time.perf_counter() - start) * 1000,
+            }
+        )
 
     def _extract_missing_module(self, stderr: str) -> str | None:
         """Extract module name from ImportError/ModuleNotFoundError."""
@@ -580,24 +595,38 @@ def _linux_sandbox_wrap(self, cmd: list[str]) -> list[str]:
 
         bwrap_cmd = [
             "bwrap",
-            "--ro-bind", "/usr", "/usr",
-            "--ro-bind", "/lib", "/lib",
-            "--ro-bind", "/bin", "/bin",
-            "--ro-bind", "/sbin", "/sbin",
+            "--ro-bind",
+            "/usr",
+            "/usr",
+            "--ro-bind",
+            "/lib",
+            "/lib",
+            "--ro-bind",
+            "/bin",
+            "/bin",
+            "--ro-bind",
+            "/sbin",
+            "/sbin",
         ]
 
         # Add /lib64 if it exists
         if os.path.exists("/lib64"):
             bwrap_cmd.extend(["--ro-bind", "/lib64", "/lib64"])
 
-        bwrap_cmd.extend([
-            "--ro-bind", self.venv_path, self.venv_path,
-            "--bind", self.temp_dir, self.temp_dir,
-            "--unshare-net",
-            "--unshare-pid",
-            "--die-with-parent",
-            "--",
-        ])
+        bwrap_cmd.extend(
+            [
+                "--ro-bind",
+                self.venv_path,
+                self.venv_path,
+                "--bind",
+                self.temp_dir,
+                self.temp_dir,
+                "--unshare-net",
+                "--unshare-pid",
+                "--die-with-parent",
+                "--",
+            ]
+        )
 
         return bwrap_cmd + cmd
 
@@ -621,11 +650,13 @@ def execute_code(self, code: str) -> REPLResult:
         total_time = time.perf_counter() - total_start
 
         # Track overhead
-        self._overhead_stats["executions"].append({
-            "total_ms": total_time * 1000,
-            "code_ms": (result.execution_time or 0) * 1000,
-            "overhead_ms": (total_time - (result.execution_time or 0)) * 1000,
-        })
+        self._overhead_stats["executions"].append(
+            {
+                "total_ms": total_time * 1000,
+                "code_ms": (result.execution_time or 0) * 1000,
+                "overhead_ms": (total_time - (result.execution_time or 0)) * 1000,
+            }
+        )
 
         return result
 
@@ -638,13 +669,15 @@ def _try_execute(self, code: str) -> REPLResult:
 
         # Build environment - inherit essential vars for macOS compatibility
         env = os.environ.copy()
-        env.update({
-            "PATH": os.path.join(self.venv_path, "bin") + ":/usr/bin:/bin",
-            "HOME": self.temp_dir,
-            "TMPDIR": self.temp_dir,
-            "PYTHONDONTWRITEBYTECODE": "1",
-            "PYTHONNOUSERSITE": "1",
-        })
+        env.update(
+            {
+                "PATH": os.path.join(self.venv_path, "bin") + ":/usr/bin:/bin",
+                "HOME": self.temp_dir,
+                "TMPDIR": self.temp_dir,
+                "PYTHONDONTWRITEBYTECODE": "1",
+                "PYTHONNOUSERSITE": "1",
+            }
+        )
 
         try:
             result = subprocess.run(
@@ -720,9 +753,7 @@ def add_context(
             context_path = os.path.join(self.temp_dir, f"context_{context_index}.txt")
             with open(context_path, "w") as f:
                 f.write(context_payload)
-            self.execute_code(
-                f"with open(r'{context_path}', 'r') as f:\n    {var_name} = f.read()"
-            )
+            self.execute_code(f"with open(r'{context_path}', 'r') as f:\n    {var_name} = f.read()")
         else:
             context_path = os.path.join(self.temp_dir, f"context_{context_index}.json")
             with open(context_path, "w") as f:
@@ -798,7 +829,9 @@ def get_overhead_summary(self) -> dict:
             "packages_installed": [i["package"] for i in installs],
             "overhead_percentage": round(
                 (total_overhead / (total_overhead + total_code_time)) * 100, 1
-            ) if total_code_time > 0 else 0,
+            )
+            if total_code_time > 0
+            else 0,
         }
 
     def print_overhead_summary(self):
diff --git a/tests/benchmarks/test_benchmarks.py b/tests/benchmarks/test_benchmarks.py
index 75d96532..1ac1c4d0 100644
--- a/tests/benchmarks/test_benchmarks.py
+++ b/tests/benchmarks/test_benchmarks.py
@@ -299,7 +299,9 @@ def custom_inference(sample):
         )
 
         # All samples should be correct (temp files weren't clobbered)
-        assert result.accuracy == 1.0, f"Some samples were clobbered: {[sr for sr in result.sample_results if not sr.is_correct]}"
+        assert result.accuracy == 1.0, (
+            f"Some samples were clobbered: {[sr for sr in result.sample_results if not sr.is_correct]}"
+        )
 
         # Verify multiple threads were used
         unique_threads = set(thread_ids.values())
@@ -398,11 +400,13 @@ def test_progress_callback_invoked(self):
         callback_invocations = []
 
         def track_callback(completed, total, sample_result, stats):
-            callback_invocations.append({
-                "completed": completed,
-                "total": total,
-                "accuracy": stats.accuracy,
-            })
+            callback_invocations.append(
+                {
+                    "completed": completed,
+                    "total": total,
+                    "accuracy": stats.accuracy,
+                }
+            )
 
         benchmark = NIAHBenchmark(context_length=1000)
         runner = BenchmarkRunner(
@@ -410,7 +414,7 @@ def track_callback(completed, total, sample_result, stats):
             progress_callback=track_callback,
         )
 
-        result = runner.run(
+        runner.run(
             benchmark,
             method="custom",
             custom_fn=lambda s: s.expected_answer,
diff --git a/tests/clients/test_huggingface.py b/tests/clients/test_huggingface.py
index 7052ca53..f9951a0f 100644
--- a/tests/clients/test_huggingface.py
+++ b/tests/clients/test_huggingface.py
@@ -19,8 +19,10 @@ class TestHuggingFaceClientUnit(unittest.TestCase):
 
     def test_init_with_api_key(self):
         """Test client initialization with explicit API key."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key", model_name="test-model")
             self.assertEqual(client.model_name, "test-model")
 
@@ -28,8 +30,10 @@ def test_init_with_base_url(self):
         """Test client initialization with base_url."""
         url = "https://example.com"
         model = "test-model"
-        with patch("rlm.clients.huggingface.InferenceClient") as MockClient, \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient") as MockClient,
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key", model_name=model, base_url=url)
             self.assertEqual(client.model_name, model)
             self.assertEqual(client.base_url, url)
@@ -39,45 +43,53 @@ def test_init_with_base_url(self):
 
             # Completion calls use model name
             mock_inst = MockClient.return_value
-            mock_inst.chat_completion.return_value.choices = [MagicMock(message=MagicMock(content="Hi"))]
+            mock_inst.chat_completion.return_value.choices = [
+                MagicMock(message=MagicMock(content="Hi"))
+            ]
 
             client.completion("Hello")
             mock_inst.chat_completion.assert_called_with(
-                messages=[{"role": "user", "content": "Hello"}],
-                model=model
+                messages=[{"role": "user", "content": "Hello"}], model=model
             )
 
     def test_init_with_url_model_name(self):
         """Test legacy behavior: model_name is URL, no base_url."""
         url = "https://example.com"
-        with patch("rlm.clients.huggingface.InferenceClient") as MockClient, \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient") as MockClient,
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key", model_name=url)
 
             # Client init with URL
             MockClient.assert_called_with(token="test-key", model=url)
 
             mock_inst = MockClient.return_value
-            mock_inst.chat_completion.return_value.choices = [MagicMock(message=MagicMock(content="Hi"))]
+            mock_inst.chat_completion.return_value.choices = [
+                MagicMock(message=MagicMock(content="Hi"))
+            ]
 
             client.completion("Hello")
             # Completion call uses None for model
             mock_inst.chat_completion.assert_called_with(
-                messages=[{"role": "user", "content": "Hello"}],
-                model=None
+                messages=[{"role": "user", "content": "Hello"}], model=None
             )
 
     def test_init_default_model(self):
         """Test client uses default model name (None if not provided)."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key")
             self.assertIsNone(client.model_name)
 
     def test_usage_tracking_initialization(self):
         """Test that usage tracking is properly initialized."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key")
             self.assertEqual(client.model_call_counts, {})
             self.assertEqual(client.model_input_tokens, {})
@@ -87,8 +99,10 @@ def test_usage_tracking_initialization(self):
 
     def test_get_usage_summary_empty(self):
         """Test usage summary when no calls have been made."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key")
             summary = client.get_usage_summary()
             self.assertIsInstance(summary, UsageSummary)
@@ -96,8 +110,10 @@ def test_get_usage_summary_empty(self):
 
     def test_get_last_usage(self):
         """Test last usage returns correct format."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key")
             client.last_prompt_tokens = 100
             client.last_completion_tokens = 50
@@ -109,16 +125,20 @@ def test_get_last_usage(self):
 
     def test_prepare_messages_string(self):
         """Test _prepare_messages with string input."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key")
             messages = client._prepare_messages("Hello world")
             self.assertEqual(messages, [{"role": "user", "content": "Hello world"}])
 
     def test_prepare_messages_list(self):
         """Test _prepare_messages with list of dicts input."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key")
             input_messages = [{"role": "user", "content": "Hello"}]
             messages = client._prepare_messages(input_messages)
@@ -126,24 +146,30 @@ def test_prepare_messages_list(self):
 
     def test_prepare_messages_invalid_type(self):
         """Test _prepare_messages raises InvalidPromptError for invalid input."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key")
             with self.assertRaises(InvalidPromptError):
                 client._prepare_messages(12345)
 
     def test_completion_requires_model(self):
         """Test completion raises ModelRequiredError when no model specified."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key", model_name=None)
             with self.assertRaises(ModelRequiredError):
                 client.completion("Hello")
 
     def test_completion_requires_model_with_base_url(self):
         """Test completion raises ModelRequiredError when using base_url without model."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"):
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+        ):
             client = HuggingFaceClient(api_key="test-key", base_url="https://example.com")
             with self.assertRaises(ModelRequiredError):
                 client.completion("Hello")
@@ -154,9 +180,11 @@ class TestHuggingFaceListModels(unittest.TestCase):
 
     def test_list_models_returns_sorted_list(self):
         """Test list_models returns sorted list of model IDs."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"), \
-             patch("rlm.clients.huggingface.list_models") as mock_list:
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+            patch("rlm.clients.huggingface.list_models") as mock_list,
+        ):
             # Mock the list_models response
             mock_model1 = MagicMock()
             mock_model1.id = "meta-llama/Llama-2-7b"
@@ -176,9 +204,11 @@ def test_list_models_returns_sorted_list(self):
 
     def test_list_models_returns_none_on_error(self):
         """Test list_models returns None when API fails."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"), \
-             patch("rlm.clients.huggingface.list_models") as mock_list:
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+            patch("rlm.clients.huggingface.list_models") as mock_list,
+        ):
             mock_list.side_effect = Exception("API Error")
 
             client = HuggingFaceClient(api_key="test-key")
@@ -188,9 +218,11 @@ def test_list_models_returns_none_on_error(self):
 
     def test_alist_models_delegates_to_list_models(self):
         """Test alist_models uses the sync list_models implementation."""
-        with patch("rlm.clients.huggingface.InferenceClient"), \
-             patch("rlm.clients.huggingface.AsyncInferenceClient"), \
-             patch("rlm.clients.huggingface.list_models") as mock_list:
+        with (
+            patch("rlm.clients.huggingface.InferenceClient"),
+            patch("rlm.clients.huggingface.AsyncInferenceClient"),
+            patch("rlm.clients.huggingface.list_models") as mock_list,
+        ):
             mock_model = MagicMock()
             mock_model.id = "test-model"
             mock_list.return_value = [mock_model]
@@ -198,6 +230,7 @@ def test_alist_models_delegates_to_list_models(self):
             client = HuggingFaceClient(api_key="test-key")
 
             import asyncio
+
             result = asyncio.get_event_loop().run_until_complete(client.alist_models())
 
             self.assertEqual(result, ["test-model"])
diff --git a/tests/environments/test_subprocess_repl.py b/tests/environments/test_subprocess_repl.py
index ffd11593..3c1793fd 100644
--- a/tests/environments/test_subprocess_repl.py
+++ b/tests/environments/test_subprocess_repl.py
@@ -9,8 +9,7 @@
 
 # Skip all tests if uv is not installed
 pytestmark = pytest.mark.skipif(
-    not shutil.which("uv"),
-    reason="uv is required for SubprocessREPL tests"
+    not shutil.which("uv"), reason="uv is required for SubprocessREPL tests"
 )
 
 
diff --git a/tests/test_exceptions.py b/tests/test_exceptions.py
index 34eace33..4d1a108c 100644
--- a/tests/test_exceptions.py
+++ b/tests/test_exceptions.py
@@ -58,7 +58,9 @@ def test_client_errors_inherit_from_client_error(self):
             InvalidPromptError,
         ]
         for exc_cls in client_exceptions:
-            assert issubclass(exc_cls, ClientError), f"{exc_cls.__name__} should inherit from ClientError"
+            assert issubclass(exc_cls, ClientError), (
+                f"{exc_cls.__name__} should inherit from ClientError"
+            )
 
     def test_environment_errors_inherit_from_environment_error(self):
         """Environment-related exceptions should inherit from EnvironmentError."""
@@ -69,7 +71,9 @@ def test_environment_errors_inherit_from_environment_error(self):
             LMQueryError,
         ]
         for exc_cls in env_exceptions:
-            assert issubclass(exc_cls, EnvironmentError), f"{exc_cls.__name__} should inherit from EnvironmentError"
+            assert issubclass(exc_cls, EnvironmentError), (
+                f"{exc_cls.__name__} should inherit from EnvironmentError"
+            )
 
     def test_registry_errors_inherit_from_registry_error(self):
         """Registry-related exceptions should inherit from RegistryError."""
@@ -79,7 +83,9 @@ def test_registry_errors_inherit_from_registry_error(self):
             ValidationError,
         ]
         for exc_cls in registry_exceptions:
-            assert issubclass(exc_cls, RegistryError), f"{exc_cls.__name__} should inherit from RegistryError"
+            assert issubclass(exc_cls, RegistryError), (
+                f"{exc_cls.__name__} should inherit from RegistryError"
+            )
 
 
 class TestClientExceptions:
@@ -110,7 +116,9 @@ def test_configuration_error_with_missing_field(self):
 
     def test_api_error_with_status_code(self):
         """APIError should store status code and response body."""
-        exc = APIError("Rate limit exceeded", status_code=429, response_body={"error": "too many requests"})
+        exc = APIError(
+            "Rate limit exceeded", status_code=429, response_body={"error": "too many requests"}
+        )
         assert "Rate limit exceeded" in str(exc)
         assert exc.status_code == 429
         assert exc.response_body == {"error": "too many requests"}
@@ -127,7 +135,9 @@ def test_variable_not_found_error_message(self):
 
     def test_code_execution_error_with_details(self):
         """CodeExecutionError should store code and stderr."""
-        exc = CodeExecutionError("Syntax error", code="print(", stderr="SyntaxError: unexpected EOF")
+        exc = CodeExecutionError(
+            "Syntax error", code="print(", stderr="SyntaxError: unexpected EOF"
+        )
         assert "Syntax error" in str(exc)
         assert exc.code == "print("
         assert exc.stderr == "SyntaxError: unexpected EOF"
diff --git a/tests/test_registry.py b/tests/test_registry.py
index 3a2588b0..2fcd279d 100644
--- a/tests/test_registry.py
+++ b/tests/test_registry.py
@@ -285,7 +285,15 @@ def test_client_config_has_env_vars_field(self):
     def test_all_backends_have_env_vars_defined(self):
         """All backends that need API keys have env_vars defined."""
         # These backends should have api_key env var
-        api_key_backends = ["openai", "openrouter", "vercel", "anthropic", "gemini", "portkey", "huggingface"]
+        api_key_backends = [
+            "openai",
+            "openrouter",
+            "vercel",
+            "anthropic",
+            "gemini",
+            "portkey",
+            "huggingface",
+        ]
         for backend in api_key_backends:
             config = CLIENT_REGISTRY[backend]
             assert "api_key" in config.env_vars, f"{backend} missing api_key env var"
diff --git a/uv.lock b/uv.lock
index 24f5b20f..335a6dd7 100644
--- a/uv.lock
+++ b/uv.lock
@@ -594,6 +594,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/9a/9a/e35b4a917281c0b8419d4207f4334c8e8c5dbf4f3f5f9ada73958d937dcc/frozenlist-1.8.0-py3-none-any.whl", hash = "sha256:0c18a16eab41e82c295618a77502e17b195883241c563b00f0aa5106fc4eaa0d", size = 13409, upload-time = "2025-10-06T05:38:16.721Z" },
 ]
 
+[[package]]
+name = "fsspec"
+version = "2026.1.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/d5/7d/5df2650c57d47c57232af5ef4b4fdbff182070421e405e0d62c6cdbfaa87/fsspec-2026.1.0.tar.gz", hash = "sha256:e987cb0496a0d81bba3a9d1cee62922fb395e7d4c3b575e57f547953334fe07b", size = 310496, upload-time = "2026-01-09T15:21:35.562Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/01/c9/97cc5aae1648dcb851958a3ddf73ccd7dbe5650d95203ecb4d7720b4cdbf/fsspec-2026.1.0-py3-none-any.whl", hash = "sha256:cb76aa913c2285a3b49bdd5fc55b1d7c708d7208126b60f2eb8194fe1b4cbdcc", size = 201838, upload-time = "2026-01-09T15:21:34.041Z" },
+]
+
 [[package]]
 name = "google-auth"
 version = "2.45.0"
@@ -669,6 +678,35 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/69/b2/119f6e6dcbd96f9069ce9a2665e0146588dc9f88f29549711853645e736a/h2-4.3.0-py3-none-any.whl", hash = "sha256:c438f029a25f7945c69e0ccf0fb951dc3f73a5f6412981daee861431b70e2bdd", size = 61779, upload-time = "2025-08-23T18:12:17.779Z" },
 ]
 
+[[package]]
+name = "hf-xet"
+version = "1.2.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/5e/6e/0f11bacf08a67f7fb5ee09740f2ca54163863b07b70d579356e9222ce5d8/hf_xet-1.2.0.tar.gz", hash = "sha256:a8c27070ca547293b6890c4bf389f713f80e8c478631432962bb7f4bc0bd7d7f", size = 506020, upload-time = "2025-10-24T19:04:32.129Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/9e/a5/85ef910a0aa034a2abcfadc360ab5ac6f6bc4e9112349bd40ca97551cff0/hf_xet-1.2.0-cp313-cp313t-macosx_10_12_x86_64.whl", hash = "sha256:ceeefcd1b7aed4956ae8499e2199607765fbd1c60510752003b6cc0b8413b649", size = 2861870, upload-time = "2025-10-24T19:04:11.422Z" },
+    { url = "https://files.pythonhosted.org/packages/ea/40/e2e0a7eb9a51fe8828ba2d47fe22a7e74914ea8a0db68a18c3aa7449c767/hf_xet-1.2.0-cp313-cp313t-macosx_11_0_arm64.whl", hash = "sha256:b70218dd548e9840224df5638fdc94bd033552963cfa97f9170829381179c813", size = 2717584, upload-time = "2025-10-24T19:04:09.586Z" },
+    { url = "https://files.pythonhosted.org/packages/a5/7d/daf7f8bc4594fdd59a8a596f9e3886133fdc68e675292218a5e4c1b7e834/hf_xet-1.2.0-cp313-cp313t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7d40b18769bb9a8bc82a9ede575ce1a44c75eb80e7375a01d76259089529b5dc", size = 3315004, upload-time = "2025-10-24T19:04:00.314Z" },
+    { url = "https://files.pythonhosted.org/packages/b1/ba/45ea2f605fbf6d81c8b21e4d970b168b18a53515923010c312c06cd83164/hf_xet-1.2.0-cp313-cp313t-manylinux_2_28_aarch64.whl", hash = "sha256:cd3a6027d59cfb60177c12d6424e31f4b5ff13d8e3a1247b3a584bf8977e6df5", size = 3222636, upload-time = "2025-10-24T19:03:58.111Z" },
+    { url = "https://files.pythonhosted.org/packages/4a/1d/04513e3cab8f29ab8c109d309ddd21a2705afab9d52f2ba1151e0c14f086/hf_xet-1.2.0-cp313-cp313t-musllinux_1_2_aarch64.whl", hash = "sha256:6de1fc44f58f6dd937956c8d304d8c2dea264c80680bcfa61ca4a15e7b76780f", size = 3408448, upload-time = "2025-10-24T19:04:20.951Z" },
+    { url = "https://files.pythonhosted.org/packages/f0/7c/60a2756d7feec7387db3a1176c632357632fbe7849fce576c5559d4520c7/hf_xet-1.2.0-cp313-cp313t-musllinux_1_2_x86_64.whl", hash = "sha256:f182f264ed2acd566c514e45da9f2119110e48a87a327ca271027904c70c5832", size = 3503401, upload-time = "2025-10-24T19:04:22.549Z" },
+    { url = "https://files.pythonhosted.org/packages/4e/64/48fffbd67fb418ab07451e4ce641a70de1c40c10a13e25325e24858ebe5a/hf_xet-1.2.0-cp313-cp313t-win_amd64.whl", hash = "sha256:293a7a3787e5c95d7be1857358a9130694a9c6021de3f27fa233f37267174382", size = 2900866, upload-time = "2025-10-24T19:04:33.461Z" },
+    { url = "https://files.pythonhosted.org/packages/e2/51/f7e2caae42f80af886db414d4e9885fac959330509089f97cccb339c6b87/hf_xet-1.2.0-cp314-cp314t-macosx_10_12_x86_64.whl", hash = "sha256:10bfab528b968c70e062607f663e21e34e2bba349e8038db546646875495179e", size = 2861861, upload-time = "2025-10-24T19:04:19.01Z" },
+    { url = "https://files.pythonhosted.org/packages/6e/1d/a641a88b69994f9371bd347f1dd35e5d1e2e2460a2e350c8d5165fc62005/hf_xet-1.2.0-cp314-cp314t-macosx_11_0_arm64.whl", hash = "sha256:2a212e842647b02eb6a911187dc878e79c4aa0aa397e88dd3b26761676e8c1f8", size = 2717699, upload-time = "2025-10-24T19:04:17.306Z" },
+    { url = "https://files.pythonhosted.org/packages/df/e0/e5e9bba7d15f0318955f7ec3f4af13f92e773fbb368c0b8008a5acbcb12f/hf_xet-1.2.0-cp314-cp314t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:30e06daccb3a7d4c065f34fc26c14c74f4653069bb2b194e7f18f17cbe9939c0", size = 3314885, upload-time = "2025-10-24T19:04:07.642Z" },
+    { url = "https://files.pythonhosted.org/packages/21/90/b7fe5ff6f2b7b8cbdf1bd56145f863c90a5807d9758a549bf3d916aa4dec/hf_xet-1.2.0-cp314-cp314t-manylinux_2_28_aarch64.whl", hash = "sha256:29c8fc913a529ec0a91867ce3d119ac1aac966e098cf49501800c870328cc090", size = 3221550, upload-time = "2025-10-24T19:04:05.55Z" },
+    { url = "https://files.pythonhosted.org/packages/6f/cb/73f276f0a7ce46cc6a6ec7d6c7d61cbfe5f2e107123d9bbd0193c355f106/hf_xet-1.2.0-cp314-cp314t-musllinux_1_2_aarch64.whl", hash = "sha256:66e159cbfcfbb29f920db2c09ed8b660eb894640d284f102ada929b6e3dc410a", size = 3408010, upload-time = "2025-10-24T19:04:28.598Z" },
+    { url = "https://files.pythonhosted.org/packages/b8/1e/d642a12caa78171f4be64f7cd9c40e3ca5279d055d0873188a58c0f5fbb9/hf_xet-1.2.0-cp314-cp314t-musllinux_1_2_x86_64.whl", hash = "sha256:9c91d5ae931510107f148874e9e2de8a16052b6f1b3ca3c1b12f15ccb491390f", size = 3503264, upload-time = "2025-10-24T19:04:30.397Z" },
+    { url = "https://files.pythonhosted.org/packages/17/b5/33764714923fa1ff922770f7ed18c2daae034d21ae6e10dbf4347c854154/hf_xet-1.2.0-cp314-cp314t-win_amd64.whl", hash = "sha256:210d577732b519ac6ede149d2f2f34049d44e8622bf14eb3d63bbcd2d4b332dc", size = 2901071, upload-time = "2025-10-24T19:04:37.463Z" },
+    { url = "https://files.pythonhosted.org/packages/96/2d/22338486473df5923a9ab7107d375dbef9173c338ebef5098ef593d2b560/hf_xet-1.2.0-cp37-abi3-macosx_10_12_x86_64.whl", hash = "sha256:46740d4ac024a7ca9b22bebf77460ff43332868b661186a8e46c227fdae01848", size = 2866099, upload-time = "2025-10-24T19:04:15.366Z" },
+    { url = "https://files.pythonhosted.org/packages/7f/8c/c5becfa53234299bc2210ba314eaaae36c2875e0045809b82e40a9544f0c/hf_xet-1.2.0-cp37-abi3-macosx_11_0_arm64.whl", hash = "sha256:27df617a076420d8845bea087f59303da8be17ed7ec0cd7ee3b9b9f579dff0e4", size = 2722178, upload-time = "2025-10-24T19:04:13.695Z" },
+    { url = "https://files.pythonhosted.org/packages/9a/92/cf3ab0b652b082e66876d08da57fcc6fa2f0e6c70dfbbafbd470bb73eb47/hf_xet-1.2.0-cp37-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3651fd5bfe0281951b988c0facbe726aa5e347b103a675f49a3fa8144c7968fd", size = 3320214, upload-time = "2025-10-24T19:04:03.596Z" },
+    { url = "https://files.pythonhosted.org/packages/46/92/3f7ec4a1b6a65bf45b059b6d4a5d38988f63e193056de2f420137e3c3244/hf_xet-1.2.0-cp37-abi3-manylinux_2_28_aarch64.whl", hash = "sha256:d06fa97c8562fb3ee7a378dd9b51e343bc5bc8190254202c9771029152f5e08c", size = 3229054, upload-time = "2025-10-24T19:04:01.949Z" },
+    { url = "https://files.pythonhosted.org/packages/0b/dd/7ac658d54b9fb7999a0ccb07ad863b413cbaf5cf172f48ebcd9497ec7263/hf_xet-1.2.0-cp37-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:4c1428c9ae73ec0939410ec73023c4f842927f39db09b063b9482dac5a3bb737", size = 3413812, upload-time = "2025-10-24T19:04:24.585Z" },
+    { url = "https://files.pythonhosted.org/packages/92/68/89ac4e5b12a9ff6286a12174c8538a5930e2ed662091dd2572bbe0a18c8a/hf_xet-1.2.0-cp37-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:a55558084c16b09b5ed32ab9ed38421e2d87cf3f1f89815764d1177081b99865", size = 3508920, upload-time = "2025-10-24T19:04:26.927Z" },
+    { url = "https://files.pythonhosted.org/packages/cb/44/870d44b30e1dcfb6a65932e3e1506c103a8a5aea9103c337e7a53180322c/hf_xet-1.2.0-cp37-abi3-win_amd64.whl", hash = "sha256:e6584a52253f72c9f52f9e549d5895ca7a471608495c4ecaa6cc73dba2b24d69", size = 2905735, upload-time = "2025-10-24T19:04:35.928Z" },
+]
+
 [[package]]
 name = "hpack"
 version = "4.1.0"
@@ -706,6 +744,27 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/2a/39/e50c7c3a983047577ee07d2a9e53faf5a69493943ec3f6a384bdc792deb2/httpx-0.28.1-py3-none-any.whl", hash = "sha256:d909fcccc110f8c7faf814ca82a9a4d816bc5a6dbfea25d6591d6985b8ba59ad", size = 73517, upload-time = "2024-12-06T15:37:21.509Z" },
 ]
 
+[[package]]
+name = "huggingface-hub"
+version = "1.3.2"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "filelock" },
+    { name = "fsspec" },
+    { name = "hf-xet", marker = "platform_machine == 'AMD64' or platform_machine == 'aarch64' or platform_machine == 'amd64' or platform_machine == 'arm64' or platform_machine == 'x86_64'" },
+    { name = "httpx" },
+    { name = "packaging" },
+    { name = "pyyaml" },
+    { name = "shellingham" },
+    { name = "tqdm" },
+    { name = "typer-slim" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/ba/d6/02d1c505e1d3364230e5fa16d2b58c8f36a39c5efe8e99bc4d03d06fd0ca/huggingface_hub-1.3.2.tar.gz", hash = "sha256:15d7902e154f04174a0816d1e9594adcf15cdad57596920a5dc70fadb5d896c7", size = 624018, upload-time = "2026-01-14T13:57:39.635Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/88/1d/acd3ef8aabb7813c6ef2f91785d855583ac5cd7c3599e5c1a1a2ed1ec2e5/huggingface_hub-1.3.2-py3-none-any.whl", hash = "sha256:b552b9562a5532102a041fa31a6966bb9de95138fc7aa578bb3703198c25d1b6", size = 534504, upload-time = "2026-01-14T13:57:37.555Z" },
+]
+
 [[package]]
 name = "hyperframe"
 version = "6.1.0"
@@ -1495,6 +1554,7 @@ source = { editable = "." }
 dependencies = [
     { name = "anthropic" },
     { name = "google-genai" },
+    { name = "huggingface-hub" },
     { name = "openai" },
     { name = "portkey-ai" },
     { name = "pytest" },
@@ -1531,6 +1591,7 @@ requires-dist = [
     { name = "dill", marker = "extra == 'modal'", specifier = ">=0.3.7" },
     { name = "dill", marker = "extra == 'prime'", specifier = ">=0.3.7" },
     { name = "google-genai", specifier = ">=1.56.0" },
+    { name = "huggingface-hub", specifier = ">=0.20.0" },
     { name = "modal", marker = "extra == 'modal'", specifier = ">=0.73.0" },
     { name = "openai", specifier = ">=2.14.0" },
     { name = "portkey-ai", specifier = ">=2.1.0" },
@@ -1741,6 +1802,19 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/e1/e4/5ebc1899d31d2b1601b32d21cfb4bba022ae6fce323d365f0448031b1660/typer-0.21.0-py3-none-any.whl", hash = "sha256:c79c01ca6b30af9fd48284058a7056ba0d3bf5cf10d0ff3d0c5b11b68c258ac6", size = 47109, upload-time = "2025-12-25T09:54:51.918Z" },
 ]
 
+[[package]]
+name = "typer-slim"
+version = "0.21.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "click" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/17/d4/064570dec6358aa9049d4708e4a10407d74c99258f8b2136bb8702303f1a/typer_slim-0.21.1.tar.gz", hash = "sha256:73495dd08c2d0940d611c5a8c04e91c2a0a98600cbd4ee19192255a233b6dbfd", size = 110478, upload-time = "2026-01-06T11:21:11.176Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/c8/0a/4aca634faf693e33004796b6cee0ae2e1dba375a800c16ab8d3eff4bb800/typer_slim-0.21.1-py3-none-any.whl", hash = "sha256:6e6c31047f171ac93cc5a973c9e617dbc5ab2bddc4d0a3135dc161b4e2020e0d", size = 47444, upload-time = "2026-01-06T11:21:12.441Z" },
+]
+
 [[package]]
 name = "types-certifi"
 version = "2021.10.8.3"