Merge remote-tracking branch 'origin/main' into litellm_mcp_client_allowlist

cookbook/litellm_proxy_server/cli_token_usage.py

@@ -3,7 +3,7 @@
 Example: Using CLI token with LiteLLM SDK
 
 This example shows how to use the CLI authentication token
-in your Python scripts after running `litellm-proxy login`.
+in your Python scripts after running `lite login`.
 """
 
 from textwrap import indent
@@ -22,7 +22,7 @@ def main():
     api_key = litellm.get_litellm_gateway_api_key()
 
     if not api_key:
-        print("❌ No CLI token found. Please run 'litellm-proxy login' first.")
+        print("❌ No CLI token found. Please run 'lite login' first.")
         return
 
     print("✅ Found CLI token.")
@@ -58,6 +58,6 @@ if __name__ == "__main__":
     main()
 
     print("\n💡 Tips:")
-    print("1. Run 'litellm-proxy login' to authenticate first")
+    print("1. Run 'lite login' to authenticate first")
     print("2. Replace 'https://your-proxy.com' with your actual proxy URL")
     print("3. The token is stored in your OS keychain, or in ~/.litellm/token.json when there is none")

litellm/proxy/_lazy_openapi_snapshot.json

@@ -38680,8 +38680,7 @@
               "required": false,
               "schema": {
                 "default": 10,
-                "maximum": 100,
-                "minimum": 1,
+                "minimum": 0,
                 "title": "Count",
                 "type": "integer"
               }
@@ -39385,8 +39384,7 @@
               "required": false,
               "schema": {
                 "default": 10,
-                "maximum": 100,
-                "minimum": 1,
+                "minimum": 0,
                 "title": "Count",
                 "type": "integer"
               }

litellm/proxy/client/cli/__init__.py

@@ -1,5 +1,5 @@
 """CLI package for LiteLLM Proxy Client."""
 
-from .main import cli
+from .main import cli, litellm_proxy_cli
 
-__all__ = ["cli"]
+__all__ = ["cli", "litellm_proxy_cli"]

litellm/proxy/client/cli/commands/encryption.py

@@ -36,8 +36,8 @@ def migrate(ctx: click.Context, check_only: bool, dry_run: bool):
     resumable; safe to re-run after an interruption.
 
     Examples:
-        litellm-proxy encryption migrate --check   # attestation scan, no writes
-        litellm-proxy encryption migrate           # perform the migration
+        lite encryption migrate --check   # attestation scan, no writes
+        lite encryption migrate           # perform the migration
     """
     client: Final = HTTPClient(ctx.obj["base_url"], ctx.obj["api_key"])
 

litellm/proxy/client/cli/main.py

@@ -168,5 +168,16 @@ cli.add_command(configure_group)
 cli.add_command(unconfigure_group)
 
 
+LITELLM_PROXY_DEPRECATION_NOTICE: Final = (
+    "The `litellm-proxy` command is deprecated and will be removed in a future release; "
+    "run `lite` instead, it takes the same commands and options."
+)
+
+
+def litellm_proxy_cli() -> None:
+    click.secho(LITELLM_PROXY_DEPRECATION_NOTICE, err=True, fg="yellow")
+    cli()
+
+
 if __name__ == "__main__":
     cli()

litellm/proxy/management_endpoints/scim/scim_v2.py

@@ -264,6 +264,8 @@ scim_router: Final = APIRouter(
     dependencies=[Depends(_premium_user_check)],
 )
 
+SCIM_MAX_PAGE_SIZE: Final = 100
+
 
 # Helper functions for common operations
 async def _get_prisma_client_or_raise_exception():
@@ -1572,12 +1574,13 @@ def _parse_scim_eq_filter(scim_filter: str) -> tuple[str, str] | None:
 )
 async def get_users(
     startIndex: int = Query(1, ge=1),
-    count: int = Query(10, ge=1, le=100),
+    count: int = Query(10, ge=0),
     filter: str | None = Query(None),
 ):
     """
     Get a list of users according to SCIM v2 protocol
     """
+    page_size: Final = min(count, SCIM_MAX_PAGE_SIZE)
     verbose_proxy_logger.debug(
         "SCIM GET USERS request: startIndex=%s count=%s filter=%s",
         startIndex,
@@ -1607,7 +1610,7 @@ async def get_users(
         users: Final[Sequence[LiteLLM_UserTable]] = await _table(UserRepository(prisma_client)).find_many(
             where=where_conditions,
             skip=(startIndex - 1),
-            take=count,
+            take=page_size,
             order={"created_at": "desc"},
         )
 
@@ -1623,7 +1626,7 @@ async def get_users(
         return SCIMListResponse(
             totalResults=total_count,
             startIndex=startIndex,
-            itemsPerPage=min(count, len(scim_users)),
+            itemsPerPage=len(scim_users),
             Resources=scim_users,
         )
 
@@ -2399,12 +2402,13 @@ class _TeamWhereConditions(TypedDict, total=False):
 )
 async def get_groups(
     startIndex: int = Query(1, ge=1),
-    count: int = Query(10, ge=1, le=100),
+    count: int = Query(10, ge=0),
     filter: str | None = Query(None),
 ):
     """
     Get a list of groups according to SCIM v2 protocol
     """
+    page_size: Final = min(count, SCIM_MAX_PAGE_SIZE)
     verbose_proxy_logger.debug(
         "SCIM GET GROUPS request: startIndex=%s count=%s filter=%s",
         startIndex,
@@ -2425,7 +2429,7 @@ async def get_groups(
         teams: Final = await _table(TeamRepository(prisma_client)).find_many(
             where=where_conditions,
             skip=(startIndex - 1),
-            take=count,
+            take=page_size,
             order={"created_at": "desc"},
         )
 
@@ -2462,7 +2466,7 @@ async def get_groups(
         return SCIMListResponse(
             totalResults=total_count,
             startIndex=startIndex,
-            itemsPerPage=min(count, len(scim_groups)),
+            itemsPerPage=len(scim_groups),
             Resources=scim_groups,
         )
 

litellm/proxy/management_endpoints/ui_sso.py

@@ -354,7 +354,7 @@ def _get_cli_sso_flow_or_raise(login_id: str | None, cache: DualCache) -> dict:
             status_code=400,
             detail=(
                 "Your litellm CLI is out of date and uses a login flow this proxy no longer supports. "
-                "Upgrade it with `pip install -U 'litellm[proxy]'` and run `litellm-proxy login` again."
+                "Upgrade it with `pip install -U 'litellm[proxy]'` and run `lite login` again."
             ),
         )
     if not _is_valid_cli_sso_login_id(login_id):
@@ -375,7 +375,7 @@ def _get_cli_sso_flow_or_raise(login_id: str | None, cache: DualCache) -> dict:
         raise HTTPException(
             status_code=400,
             detail=(
-                "CLI login session not found or expired. Run `litellm-proxy login` again. "
+                "CLI login session not found or expired. Run `lite login` again. "
                 "If this happens immediately after starting a login, the proxy is likely running multiple "
                 "replicas without a shared cache; configure a Redis cache "
                 "so every replica can see the login session."

pyproject.toml

@@ -174,7 +174,7 @@ proxy-runtime = [
 [project.scripts]
 litellm = "litellm:run_server"
 lite = "litellm.proxy.client.cli:cli"
-litellm-proxy = "litellm.proxy.client.cli:cli"
+litellm-proxy = "litellm.proxy.client.cli:litellm_proxy_cli"
 
 [dependency-groups]
 dev = [

tests/otel_tests/test_e2e_budgeting.py

@@ -367,7 +367,7 @@ async def obtain_cli_sso_token_via_poll_flow(
     models: list[str],
 ) -> str:
     """
-    Obtain a CLI SSO JWT through the same HTTP flow as `litellm-proxy login`:
+    Obtain a CLI SSO JWT through the same HTTP flow as `lite login`:
     /sso/cli/start -> (SSO callback) -> /sso/cli/complete -> /sso/cli/poll.
 
     When the proxy SSO session cache is not shared with the test runner (otel CI
@@ -551,7 +551,7 @@ async def test_team_budget_enforcement():
 @pytest.mark.asyncio
 async def test_team_budget_enforcement_cli_sso_token():
     """
-    Team budget enforcement for CLI SSO session tokens (litellm-proxy login JWT).
+    Team budget enforcement for CLI SSO session tokens (lite login JWT).
 
     1. Create team with a tiny max_budget and a user on that team
     2. Obtain a CLI SSO JWT (HTTP poll flow when Redis is shared, else mint)

tests/test_litellm/proxy/client/cli/test_encryption_commands.py

@@ -1,4 +1,4 @@
-"""CLI tests for the ``litellm-proxy encryption migrate`` command.
+"""CLI tests for the ``lite encryption migrate`` command.
 
 The HTTP client is mocked, so these assert the command's request routing (GET
 check vs POST migrate, dry-run param) and its residual-state messaging without a

tests/test_litellm/proxy/client/cli/test_global_options.py

@@ -1,7 +1,9 @@
 # stdlib imports
 import json
 import os
+import sys
 from pathlib import Path
+from typing import Final
 from unittest.mock import Mock, patch
 
 import pytest
@@ -9,7 +11,8 @@ from click.testing import CliRunner
 
 import litellm.proxy.client.cli
 from litellm._version import version as litellm_version
-from litellm.proxy.client.cli import cli
+from litellm.proxy.client.cli import cli, litellm_proxy_cli
+from litellm.proxy.client.cli.main import LITELLM_PROXY_DEPRECATION_NOTICE
 
 
 @pytest.fixture
@@ -234,3 +237,32 @@ def test_version_flag_never_sends_api_key_to_unnamed_server(cli_runner, isolated
     assert all(url.startswith("https://flag-proxy.example.com") for url in requested_urls)
     sent_keys = [call.kwargs["headers"].get("Authorization") for call in mock_request.call_args_list]
     assert sent_keys == ["Bearer sk-intended-for-flag-proxy"] * len(requested_urls)
+
+
+def test_litellm_proxy_entrypoint_prints_deprecation_notice_on_stderr_and_still_runs(monkeypatch, capsys, requests_mock):
+    requests_mock.get("http://localhost:4000/health/readiness", json={"litellm_version": "1.2.3"})
+    monkeypatch.setattr(sys, "argv", ["litellm-proxy", "--version"])
+    monkeypatch.setenv("LITELLM_PROXY_URL", "http://localhost:4000")
+    with pytest.raises(SystemExit) as exit_info:
+        litellm_proxy_cli()
+
+    captured: Final = capsys.readouterr()
+    assert exit_info.value.code == 0
+    assert captured.err.strip() == LITELLM_PROXY_DEPRECATION_NOTICE
+    assert f"LiteLLM Proxy CLI Version: {litellm_version}" in captured.out
+    assert "LiteLLM Proxy Server Version: 1.2.3" in captured.out
+    assert "deprecated" not in captured.out
+
+
+def test_lite_entrypoint_prints_nothing_on_stderr(monkeypatch, capsys, requests_mock):
+    requests_mock.get("http://localhost:4000/health/readiness", json={"litellm_version": "1.2.3"})
+    monkeypatch.setattr(sys, "argv", ["lite", "--version"])
+    monkeypatch.setenv("LITELLM_PROXY_URL", "http://localhost:4000")
+    with pytest.raises(SystemExit) as exit_info:
+        cli()
+
+    captured: Final = capsys.readouterr()
+    assert exit_info.value.code == 0
+    assert "LiteLLM Proxy Server Version: 1.2.3" in captured.out
+    assert f"LiteLLM Proxy CLI Version: {litellm_version}" in captured.out
+    assert captured.err == ""

tests/test_litellm/proxy/management_endpoints/scim/test_scim_v2_endpoints.py

@@ -7,7 +7,8 @@ from typing import Final
 from unittest.mock import AsyncMock, MagicMock, call
 
 import pytest
-from fastapi import HTTPException
+from fastapi import FastAPI, HTTPException
+from httpx import ASGITransport, AsyncClient
 from pytest_mock import MockerFixture
 
 from litellm.proxy._types import (
@@ -31,6 +32,7 @@ from litellm.proxy.management_endpoints.scim.scim_v2 import (
     _handle_group_membership_changes,
     _handle_team_membership_changes,
     _parse_member_entries,
+    _premium_user_check,
     _process_group_patch_operations,
     _recompute_scim_member_roles,
     _resolve_group_member_ids,
@@ -45,8 +47,10 @@ from litellm.proxy.management_endpoints.scim.scim_v2 import (
     patch_group,
     patch_team_membership,
     patch_user,
+    scim_router,
     update_group,
     update_user,
+    user_api_key_auth,
 )
 from litellm.types.proxy.management_endpoints.scim_v2 import (
     SCIM_ENTERPRISE_USER_SCHEMA,
@@ -484,6 +488,48 @@ async def test_scim_create_user_respects_default_role_set_via_ui(mocker, monkeyp
     )
 
 
+@pytest.fixture
+def scim_test_client():
+    """An in-process SCIM application with authorization dependencies bypassed."""
+    app = FastAPI()
+    app.dependency_overrides[_premium_user_check] = lambda: None
+    app.dependency_overrides[user_api_key_auth] = lambda: UserAPIKeyAuth(user_role=LitellmUserRoles.PROXY_ADMIN)
+    app.include_router(scim_router)
+    return AsyncClient(transport=ASGITransport(app=app), base_url="http://test")
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize("endpoint", ["Users", "Groups"])
+@pytest.mark.parametrize(("requested_count", "effective_count"), [(0, 0), (200, 100), (1000, 100)])
+async def test_scim_collection_endpoints_clamp_requested_page_size(
+    scim_test_client, endpoint, requested_count, effective_count, mocker
+):
+    """SCIM list endpoints accept zero and cap larger client page requests."""
+    mock_prisma_client = MagicMock()
+    mock_prisma_client.db = MagicMock()
+    table = MagicMock()
+    table.find_many = AsyncMock(return_value=[])
+    table.count = AsyncMock(return_value=0)
+    mock_prisma_client.db.litellm_usertable = table
+    mock_prisma_client.db.litellm_teamtable = table
+    mocker.patch(  # test-quality-ok: HTTP validation requires an in-memory database boundary.
+        "litellm.proxy.management_endpoints.scim.scim_v2._get_prisma_client_or_raise_exception",
+        AsyncMock(return_value=mock_prisma_client),
+    )
+
+    async with scim_test_client as client:
+        response = await client.get(f"/scim/v2/{endpoint}?startIndex=1&count={requested_count}")
+
+    assert response.status_code == 200
+    table.find_many.assert_awaited_once_with(
+        where={},
+        skip=0,
+        take=effective_count,
+        order={"created_at": "desc"},
+    )
+    assert response.json()["itemsPerPage"] == 0
+
+
 @pytest.mark.asyncio
 async def test_get_users_filters_username_by_exposed_scim_username_for_okta(mocker):
     """