Fix useCreateConversationAndSubscribeMultiple to handle full lifecycle automatically

- Restructured hook to automatically poll and subscribe to conversations after creation
- Added immediate 'starting' toast for better UX during polling period
- Fixed React hooks rules violations by using useEffect with Promise.all instead of loops
- Removed unused functions and cleaned up imports
- Hook now handles: create → show starting toast → poll until ready → subscribe → cleanup

Co-authored-by: OpenHands <openhands@all-hands.dev>

.github/workflows/py-tests.yml

@@ -48,11 +48,11 @@ jobs:
       - name: Build Environment
         run: make build
       - name: Run Unit Tests
-        run: poetry run pytest --forked -n auto -svv ./tests/unit
+        run: PYTHONPATH=".:$PYTHONPATH" poetry run pytest --forked -n auto -svv ./tests/unit
       - name: Run Runtime Tests with CLIRuntime
-        run: TEST_RUNTIME=cli poetry run pytest -svv tests/runtime/test_bash.py
+        run: PYTHONPATH=".:$PYTHONPATH" TEST_RUNTIME=cli poetry run pytest -svv tests/runtime/test_bash.py
       - name: Run E2E Tests
-        run: poetry run pytest -svv tests/e2e
+        run: PYTHONPATH=".:$PYTHONPATH" poetry run pytest -svv tests/e2e
 
   # Run specific Windows python tests
   test-on-windows:
@@ -77,9 +77,11 @@ jobs:
       - name: Run Windows unit tests
         run: poetry run pytest -svv tests/unit/test_windows_bash.py
         env:
+          PYTHONPATH: ".;$env:PYTHONPATH"
           DEBUG: "1"
       - name: Run Windows runtime tests with LocalRuntime
         run: $env:TEST_RUNTIME="local"; poetry run pytest -svv tests/runtime/test_bash.py
         env:
+          PYTHONPATH: ".;$env:PYTHONPATH"
           TEST_RUNTIME: local
           DEBUG: "1"

evaluation/benchmarks/swe_bench/README.md

@@ -183,24 +183,7 @@ The final results will be saved to `evaluation/evaluation_outputs/outputs/swe_be
 - `report.json`: a JSON file that contains keys like `"resolved_ids"` pointing to instance IDs that are resolved by the agent.
 - `logs/`: a directory of test logs
 
-### Run evaluation with `RemoteRuntime`
 
-OpenHands Remote Runtime is currently in beta (read [here](https://runtime.all-hands.dev/) for more details), it allows you to run rollout in parallel in the cloud, so you don't need a powerful machine to run evaluation.
-Fill out [this form](https://docs.google.com/forms/d/e/1FAIpQLSckVz_JFwg2_mOxNZjCtr7aoBFI2Mwdan3f75J_TrdMS1JV2g/viewform) to apply if you want to try this out!
-
-```bash
-./evaluation/benchmarks/swe_bench/scripts/eval_infer_remote.sh [output.jsonl filepath] [num_workers]
-
-# Example - This evaluates patches generated by CodeActAgent on Llama-3.1-70B-Instruct-Turbo on "princeton-nlp/SWE-bench_Lite"'s test set, with 16 number of workers running in parallel
-ALLHANDS_API_KEY="YOUR-API-KEY" RUNTIME=remote SANDBOX_REMOTE_RUNTIME_API_URL="https://runtime.eval.all-hands.dev" EVAL_DOCKER_IMAGE_PREFIX="us-central1-docker.pkg.dev/evaluation-092424/swe-bench-images" \
-evaluation/benchmarks/swe_bench/scripts/eval_infer_remote.sh evaluation/evaluation_outputs/outputs/swe-bench-lite/CodeActAgent/Llama-3.1-70B-Instruct-Turbo_maxiter_100_N_v1.9-no-hint/output.jsonl 16 "princeton-nlp/SWE-bench_Lite" "test"
-```
-
-To clean-up all existing runtimes that you've already started, run:
-
-```bash
-ALLHANDS_API_KEY="YOUR-API-KEY" ./evaluation/utils/scripts/cleanup_remote_runtime.sh
-```
 
 ## SWT-Bench Evaluation
 

evaluation/benchmarks/swe_bench/scripts/eval_infer_remote.sh

@@ -1,46 +0,0 @@
-#!/usr/bin/env bash
-set -eo pipefail
-
-INPUT_FILE=$1
-NUM_WORKERS=$2
-DATASET=$3
-SPLIT=$4
-
-if [ -z "$INPUT_FILE" ]; then
-  echo "INPUT_FILE not specified (should be a path to a jsonl file)"
-  exit 1
-fi
-
-if [ -z "$DATASET" ]; then
-  echo "DATASET not specified, use default princeton-nlp/SWE-bench_Lite"
-  DATASET="princeton-nlp/SWE-bench_Lite"
-fi
-
-if [ -z "$SPLIT" ]; then
-  echo "SPLIT not specified, use default test"
-  SPLIT="test"
-fi
-
-if [ -z "$NUM_WORKERS" ]; then
-  echo "NUM_WORKERS not specified, use default 1"
-  NUM_WORKERS=1
-fi
-
-echo "... Evaluating on $INPUT_FILE ..."
-
-COMMAND="poetry run python evaluation/benchmarks/swe_bench/eval_infer.py \
-  --eval-num-workers $NUM_WORKERS \
-  --input-file $INPUT_FILE \
-  --dataset $DATASET \
-  --split $SPLIT"
-
-if [ -n "$EVAL_LIMIT" ]; then
-  echo "EVAL_LIMIT: $EVAL_LIMIT"
-  COMMAND="$COMMAND --eval-n-limit $EVAL_LIMIT"
-fi
-
-# Run the command
-eval $COMMAND
-
-# update the output with evaluation results
-poetry run python evaluation/benchmarks/swe_bench/scripts/eval/update_output_with_eval.py $INPUT_FILE

evaluation/benchmarks/testgeneval/NOTES.md

@@ -5,8 +5,7 @@ pynguin_ids = ['pydata__xarray-6548-16541', 'pydata__xarray-7003-16557', 'pydata
 ids = ['pydata__xarray-3114-16452', 'pydata__xarray-3151-16453', 'pydata__xarray-3156-16454', 'pydata__xarray-3239-16456', 'pydata__xarray-3239-16457', 'pydata__xarray-3239-16458', 'pydata__xarray-3302-16459', 'pydata__xarray-3364-16461', 'pydata__xarray-3677-16471', 'pydata__xarray-3905-16478', 'pydata__xarray-4182-16484', 'pydata__xarray-4248-16486', 'pydata__xarray-4339-16487', 'pydata__xarray-4419-16488', 'pydata__xarray-4629-16492', 'pydata__xarray-4750-16496', 'pydata__xarray-4802-16505', 'pydata__xarray-4966-16515', 'pydata__xarray-4994-16516', 'pydata__xarray-5033-16517', 'pydata__xarray-5126-16518', 'pydata__xarray-5126-16519', 'pydata__xarray-5131-16520', 'pydata__xarray-5365-16529', 'pydata__xarray-5455-16530', 'pydata__xarray-5662-16532', 'pydata__xarray-5731-16534', 'pydata__xarray-6135-16535', 'pydata__xarray-6135-16536', 'pydata__xarray-6386-16537', 'pydata__xarray-6394-16538', 'pydata__xarray-6400-16539', 'pydata__xarray-6461-16540', 'pydata__xarray-6548-16541', 'pydata__xarray-6599-16543', 'pydata__xarray-6601-16544', 'pydata__xarray-6882-16548', 'pydata__xarray-6889-16549', 'pydata__xarray-7003-16557', 'pydata__xarray-7147-16571', 'pydata__xarray-7150-16572', 'pydata__xarray-7203-16577', 'pydata__xarray-7229-16578', 'pydata__xarray-7393-16581', 'pydata__xarray-7400-16582']
 
 
-Command eval (our approach):
-poetry run ./evaluation/benchmarks/testgeneval/scripts/eval_infer_remote.sh evaluation/evaluation_outputs/outputs/kjain14__testgeneval-test/CodeActAgent/gpt-4o_maxiter_25_N_v0.20.0-no-hint-run_1/output.jsonl 10 kjain14/testgeneval test true
+
 
 Command run (our approach):
 ./evaluation/benchmarks/testgeneval/scripts/run_infer.sh llm.eval_gpt HEAD CodeActAgent -1 25 10 kjain14/testgeneval test 1 ../TestGenEval/results/testgeneval/preds/gpt-4o-2024-08-06__testgeneval__0.2__test.jsonl

frontend/__tests__/utils/extract-model-and-provider.test.ts

@@ -82,5 +82,11 @@ describe("extractModelAndProvider", () => {
       model: "claude-opus-4-20250514",
       separator: "/",
     });
+
+    expect(extractModelAndProvider("claude-opus-4-1-20250805")).toEqual({
+      provider: "anthropic",
+      model: "claude-opus-4-1-20250805",
+      separator: "/",
+    });
   });
 });

frontend/index.html

@@ -1,44 +0,0 @@
-<!DOCTYPE html>
-<html lang="en">
-  <head>
-    <meta charset="utf-8" />
-    <link rel="apple-touch-icon" sizes="180x180" href="/apple-touch-icon.png">
-    <link rel="icon" type="image/png" sizes="32x32" href="/favicon-32x32.png">
-    <link rel="icon" type="image/png" sizes="16x16" href="/favicon-16x16.png">
-    <link rel="manifest" href="/site.webmanifest">
-    <link rel="mask-icon" href="/safari-pinned-tab.svg" color="#5bbad5">
-    <meta name="msapplication-TileColor" content="#da532c">
-    <meta name="theme-color" content="#ffffff">
-    <meta name="viewport" content="width=device-width, initial-scale=1" />
-    <meta name="theme-color" content="#000000" />
-    <meta
-      name="description"
-      content="OpenHands: Code Less, Make More"
-    />
-    <!--
-      Notice the use of %PUBLIC_URL% in the tags above.
-      It will be replaced with the URL of the `public` folder during the build.
-      Only files inside the `public` folder can be referenced from the HTML.
-
-      Unlike "/favicon.ico" or "favicon.ico", "%PUBLIC_URL%/favicon.ico" will
-      work correctly both with client-side routing and a non-root public URL.
-      Learn how to configure a non-root public URL by running `npm run build`.
-    -->
-    <title>OpenHands</title>
-  </head>
-  <body>
-    <noscript>You need to enable JavaScript to run this app.</noscript>
-    <div id="root"></div>
-    <!--
-      This HTML file is a template.
-      If you open it directly in the browser, you will see an empty page.
-
-      You can add webfonts, meta tags, or analytics to this file.
-      The build step will place the bundled scripts into the <body> tag.
-
-      To begin the development, run `npm start` or `yarn start`.
-      To create a production bundle, use `npm run build` or `yarn build`.
-    -->
-    <script type="module" src="/src/index.tsx"></script>
-  </body>
-</html>

frontend/src/hooks/use-create-conversation-and-subscribe-multiple.ts

@@ -1,9 +1,10 @@
 import React from "react";
+import toast from "react-hot-toast";
 import { useCreateConversation } from "./mutation/use-create-conversation";
-import { useUserProviders } from "./use-user-providers";
 import { useConversationSubscriptions } from "#/context/conversation-subscriptions-provider";
 import { Provider } from "#/types/settings";
 import { CreateMicroagent } from "#/api/open-hands.types";
+import { TOAST_OPTIONS } from "#/utils/custom-toast-handlers";
 import OpenHands from "#/api/open-hands";
 
 /**
@@ -11,16 +12,87 @@ import OpenHands from "#/api/open-hands";
  * This extends the functionality of useCreateConversationAndSubscribe to allow subscribing to
  * multiple conversations simultaneously.
  */
+
 export const useCreateConversationAndSubscribeMultiple = () => {
   const { mutate: createConversation, isPending } = useCreateConversation();
-  const { providers } = useUserProviders();
   const {
-    subscribeToConversation,
     unsubscribeFromConversation,
     isSubscribedToConversation,
     activeConversationIds,
+    subscribeToConversation,
   } = useConversationSubscriptions();
 
+  // Track conversations that are being set up with their callbacks
+  const [pendingConversations, setPendingConversations] = React.useState<
+    Map<
+      string,
+      {
+        onEventCallback?: (event: unknown, conversationId: string) => void;
+        repositoryName: string;
+      }
+    >
+  >(new Map());
+
+  // Handle polling and subscription for pending conversations
+  React.useEffect(() => {
+    const handleConversationPolling = async () => {
+      const conversationsToProcess = Array.from(pendingConversations.entries());
+
+      await Promise.all(
+        conversationsToProcess.map(async ([conversationId, config]) => {
+          try {
+            const conversation =
+              await OpenHands.getConversation(conversationId);
+
+            if (
+              conversation?.status === "RUNNING" &&
+              conversation.runtime_status
+            ) {
+              // Conversation is ready, subscribe to it
+              let baseUrl = "";
+              if (conversation.url && !conversation.url.startsWith("/")) {
+                baseUrl = new URL(conversation.url).host;
+              } else {
+                baseUrl =
+                  (import.meta.env.VITE_BACKEND_BASE_URL as
+                    | string
+                    | undefined) || window?.location.host;
+              }
+
+              subscribeToConversation({
+                conversationId,
+                sessionApiKey: conversation.session_api_key,
+                providersSet: [], // Empty array since we don't need providers for subscription
+                baseUrl,
+                onEvent: config.onEventCallback,
+              });
+
+              // Remove from pending when subscription is established
+              setPendingConversations((prev) => {
+                const newMap = new Map(prev);
+                newMap.delete(conversationId);
+                return newMap;
+              });
+            }
+          } catch (error) {
+            // Remove failed conversation from pending
+            setPendingConversations((prev) => {
+              const newMap = new Map(prev);
+              newMap.delete(conversationId);
+              return newMap;
+            });
+          }
+        }),
+      );
+    };
+
+    if (pendingConversations.size > 0) {
+      const interval = setInterval(handleConversationPolling, 1000);
+      return () => clearInterval(interval);
+    }
+    return undefined;
+  }, [pendingConversations, subscribeToConversation]);
+
   const createConversationAndSubscribe = React.useCallback(
     ({
       query,
@@ -49,77 +121,33 @@ export const useCreateConversationAndSubscribeMultiple = () => {
           createMicroagent,
         },
         {
-          onSuccess: async (data) => {
-            try {
-              // NOTE: createConversation returns ConversationResponse (no url/session_api_key)
-              // but we need the full Conversation object for WebSocket connection.
-              // Wait for conversation to be fully loaded to get proper url and session_api_key
-              const conversation = await OpenHands.getConversation(
-                data.conversation_id,
-              );
-              if (!conversation) {
-                // eslint-disable-next-line no-console
-                console.error("Failed to load conversation after creation");
-                return;
-              }
-
-              let baseUrl = "";
-              if (conversation.url && !conversation.url.startsWith("/")) {
-                baseUrl = new URL(conversation.url).host;
-              } else {
-                baseUrl =
-                  (import.meta.env.VITE_BACKEND_BASE_URL as
-                    | string
-                    | undefined) || window?.location.host;
-              }
-
-              // Subscribe to the conversation using the loaded conversation data
-              subscribeToConversation({
-                conversationId: conversation.conversation_id,
-                sessionApiKey: conversation.session_api_key,
-                providersSet: providers,
-                baseUrl,
-                onEvent: onEventCallback,
+          onSuccess: (data) => {
+            // Add to pending conversations for polling
+            setPendingConversations((prev) => {
+              const newMap = new Map(prev);
+              newMap.set(data.conversation_id, {
+                onEventCallback,
+                repositoryName: repository.name,
               });
+              return newMap;
+            });
 
-              // Call the success callback if provided
-              if (onSuccessCallback) {
-                onSuccessCallback(data.conversation_id);
-              }
-            } catch (error) {
-              // eslint-disable-next-line no-console
-              console.error(
-                "Error loading conversation for WebSocket connection:",
-                error,
-              );
-              // Fallback to original behavior if fetching conversation fails
-              let baseUrl = "";
-              if (data?.url && !data.url.startsWith("/")) {
-                baseUrl = new URL(data.url).host;
-              } else {
-                baseUrl =
-                  (import.meta.env.VITE_BACKEND_BASE_URL as
-                    | string
-                    | undefined) || window?.location.host;
-              }
+            // Show immediate "starting" toast to give user feedback
+            toast(`Starting conversation for ${repository.name}...`, {
+              ...TOAST_OPTIONS,
+              id: `starting-${data.conversation_id}`,
+              duration: 10000, // Longer duration since this will be replaced by the runtime status toast
+            });
 
-              subscribeToConversation({
-                conversationId: data.conversation_id,
-                sessionApiKey: data.session_api_key,
-                providersSet: providers,
-                baseUrl,
-                onEvent: onEventCallback,
-              });
-
-              if (onSuccessCallback) {
-                onSuccessCallback(data.conversation_id);
-              }
+            // Call the success callback immediately (conversation created)
+            if (onSuccessCallback) {
+              onSuccessCallback(data.conversation_id);
             }
           },
         },
       );
     },
-    [createConversation, subscribeToConversation, providers],
+    [createConversation],
   );
 
   return {

frontend/src/utils/verified-models.ts

@@ -14,6 +14,7 @@ export const VERIFIED_MODELS = [
   "claude-3-7-sonnet-20250219",
   "claude-sonnet-4-20250514",
   "claude-opus-4-20250514",
+  "claude-opus-4-1-20250805",
   "gemini-2.5-pro",
   "o4-mini",
   "deepseek-chat",
@@ -47,6 +48,7 @@ export const VERIFIED_ANTHROPIC_MODELS = [
   "claude-3-7-sonnet-20250219",
   "claude-sonnet-4-20250514",
   "claude-opus-4-20250514",
+  "claude-opus-4-1-20250805",
 ];
 
 // LiteLLM does not return the compatible Mistral models with the provider, so we list them here to set them ourselves
@@ -62,6 +64,7 @@ export const VERIFIED_MISTRAL_MODELS = [
 export const VERIFIED_OPENHANDS_MODELS = [
   "claude-sonnet-4-20250514",
   "claude-opus-4-20250514",
+  "claude-opus-4-1-20250805",
   "gemini-2.5-pro",
   "o3",
   "o4-mini",

openhands/cli/main.py

@@ -129,12 +129,13 @@ async def run_session(
     conversation_instructions: str | None = None,
     session_name: str | None = None,
     skip_banner: bool = False,
+    conversation_id: str | None = None,
 ) -> bool:
     reload_microagents = False
     new_session_requested = False
     exit_reason = ExitReason.INTENTIONAL
 
-    sid = generate_sid(config, session_name)
+    sid = conversation_id or generate_sid(config, session_name)
     is_loaded = asyncio.Event()
     is_paused = asyncio.Event()  # Event to track agent pause requests
     always_confirm_mode = False  # Flag to enable always confirm mode
@@ -705,6 +706,7 @@ After reviewing the file, please ask the user what they would like to do with it
         task_str,
         session_name=args.name,
         skip_banner=banner_shown,
+        conversation_id=args.conversation,
     )
 
     # If a new session was requested, run it

openhands/cli/settings.py

@@ -27,7 +27,7 @@ from openhands.core.config.condenser_config import (
     CondenserPipelineConfig,
     ConversationWindowCondenserConfig,
 )
-from openhands.core.config.utils import OH_DEFAULT_AGENT
+from openhands.core.config.config_utils import OH_DEFAULT_AGENT
 from openhands.memory.condenser.impl.llm_summarizing_condenser import (
     LLMSummarizingCondenserConfig,
 )

openhands/cli/utils.py

@@ -164,6 +164,7 @@ VERIFIED_OPENAI_MODELS = [
 VERIFIED_ANTHROPIC_MODELS = [
     'claude-sonnet-4-20250514',
     'claude-opus-4-20250514',
+    'claude-opus-4-1-20250805',
     'claude-3-7-sonnet-20250219',
     'claude-3-sonnet-20240229',
     'claude-3-opus-20240229',
@@ -184,6 +185,7 @@ VERIFIED_MISTRAL_MODELS = [
 VERIFIED_OPENHANDS_MODELS = [
     'claude-sonnet-4-20250514',
     'claude-opus-4-20250514',
+    'claude-opus-4-1-20250805',
     'devstral-small-2507',
     'devstral-medium-2507',
     'o3',

openhands/core/config/utils.py

@@ -20,10 +20,6 @@ from openhands.core.config.condenser_config import (
     condenser_config_from_toml_section,
     create_condenser_config,
 )
-from openhands.core.config.config_utils import (
-    OH_DEFAULT_AGENT,
-    OH_MAX_ITERATIONS,
-)
 from openhands.core.config.extended_config import ExtendedConfig
 from openhands.core.config.kubernetes_config import KubernetesConfig
 from openhands.core.config.llm_config import LLMConfig
@@ -712,14 +708,14 @@ def get_parser() -> argparse.ArgumentParser:
     parser.add_argument(
         '-c',
         '--agent-cls',
-        default=OH_DEFAULT_AGENT,
+        default=None,
         type=str,
         help='Name of the default agent to use',
     )
     parser.add_argument(
         '-i',
         '--max-iterations',
-        default=OH_MAX_ITERATIONS,
+        default=None,
         type=int,
         help='The maximum number of iterations to run the agent',
     )
@@ -774,6 +770,12 @@ def get_parser() -> argparse.ArgumentParser:
         type=str,
         default='',
     )
+    parser.add_argument(
+        '--conversation',
+        help='The conversation id to continue',
+        type=str,
+        default=None,
+    )
     parser.add_argument(
         '--eval-ids',
         default=None,

openhands/llm/llm.py

@@ -63,6 +63,7 @@ CACHE_PROMPT_SUPPORTED_MODELS = [
     'claude-sonnet-4-20250514',
     'claude-sonnet-4',
     'claude-opus-4-20250514',
+    'claude-opus-4-1-20250805',
 ]
 
 # function calling supporting models
@@ -77,6 +78,7 @@ FUNCTION_CALLING_SUPPORTED_MODELS = [
     'claude-sonnet-4-20250514',
     'claude-sonnet-4',
     'claude-opus-4-20250514',
+    'claude-opus-4-1-20250805',
     'gpt-4o-mini',
     'gpt-4o',
     'o1-2024-12-17',

package-lock.json

@@ -1,6 +0,0 @@
-{
-  "name": "OpenHands",
-  "lockfileVersion": 3,
-  "requires": true,
-  "packages": {}
-}

tests/unit/test_arg_parser.py

@@ -1,6 +1,6 @@
 import pytest
 
-from openhands.core.config import OH_DEFAULT_AGENT, OH_MAX_ITERATIONS, get_parser
+from openhands.core.config import get_parser
 
 
 def test_parser_default_values():
@@ -10,8 +10,8 @@ def test_parser_default_values():
     assert args.directory is None
     assert args.task == ''
     assert args.file is None
-    assert args.agent_cls == OH_DEFAULT_AGENT
-    assert args.max_iterations == OH_MAX_ITERATIONS
+    assert args.agent_cls is None
+    assert args.max_iterations is None
     assert args.max_budget_per_task is None
     assert args.eval_output_dir == 'evaluation/evaluation_outputs/outputs'
     assert args.eval_n_limit is None
@@ -139,13 +139,14 @@ def test_help_message(capsys):
         '--selected-repo SELECTED_REPO',
         '--override-cli-mode OVERRIDE_CLI_MODE',
         '--log-level LOG_LEVEL',
+        '--conversation CONVERSATION',
     ]
 
     for element in expected_elements:
         assert element in help_output, f"Expected '{element}' to be in the help message"
 
     option_count = help_output.count('  -')
-    assert option_count == 21, f'Expected 21 options, found {option_count}'
+    assert option_count == 22, f'Expected 22 options, found {option_count}'
 
 
 def test_selected_repo_format():

tests/unit/test_cli.py

@@ -359,6 +359,7 @@ async def test_main_without_task(
     mock_args.llm_config = None
     mock_args.name = None
     mock_args.file = None
+    mock_args.conversation = None
     mock_parse_args.return_value = mock_args
 
     # Mock config
@@ -412,6 +413,7 @@ async def test_main_without_task(
         None,
         session_name=None,
         skip_banner=False,
+        conversation_id=None,
     )
 
 
@@ -553,6 +555,7 @@ async def test_main_with_session_name_passes_name_to_run_session(
     mock_args.llm_config = None
     mock_args.name = test_session_name  # Set the session name
     mock_args.file = None
+    mock_args.conversation = None
     mock_parse_args.return_value = mock_args
 
     # Mock config
@@ -606,6 +609,7 @@ async def test_main_with_session_name_passes_name_to_run_session(
         None,
         session_name=test_session_name,
         skip_banner=False,
+        conversation_id=None,
     )
 
 

tests/unit/test_config_precedence.py

@@ -3,6 +3,8 @@ from unittest.mock import MagicMock, patch
 import pytest
 
 from openhands.core.config import (
+    OH_DEFAULT_AGENT,
+    OH_MAX_ITERATIONS,
     OpenHandsConfig,
     get_llm_config_arg,
     setup_config_from_args,
@@ -308,3 +310,74 @@ def test_cli_settings_json_not_override_config_toml(
     # Verify that settings.json did not override config.toml
     assert test_llm_config.model == 'config-toml-model'
     assert test_llm_config.api_key == 'config-toml-api-key'
+
+
+def test_default_values_applied_when_none():
+    """Test that default values are applied when config values are None."""
+
+    # Create mock args with None values for agent_cls and max_iterations
+    mock_args = MagicMock()
+    mock_args.config_file = None
+    mock_args.llm_config = None
+    mock_args.agent_cls = None
+    mock_args.max_iterations = None
+
+    # Load config
+    with patch(
+        'openhands.core.config.utils.load_openhands_config',
+        return_value=OpenHandsConfig(),
+    ):
+        config = setup_config_from_args(mock_args)
+
+    # Verify they match the expected defaults
+    assert config.default_agent == OH_DEFAULT_AGENT
+    assert config.max_iterations == OH_MAX_ITERATIONS
+
+
+def test_cli_args_override_defaults():
+    """Test that CLI arguments override default values."""
+
+    # Create mock args with custom values
+    mock_args = MagicMock()
+    mock_args.config_file = None
+    mock_args.llm_config = None
+    mock_args.agent_cls = 'CustomAgent'
+    mock_args.max_iterations = 50
+
+    # Load config
+    with patch(
+        'openhands.core.config.utils.load_openhands_config',
+        return_value=OpenHandsConfig(),
+    ):
+        config = setup_config_from_args(mock_args)
+
+    # Verify custom values are used instead of defaults
+    assert config.default_agent == 'CustomAgent'
+    assert config.max_iterations == 50
+
+
+def test_cli_args_none_uses_config_toml_values():
+    """Test that when CLI args agent_cls and max_iterations are None, config.toml values are used."""
+
+    # Create mock args with None values for agent_cls and max_iterations
+    mock_args = MagicMock()
+    mock_args.config_file = None
+    mock_args.llm_config = None
+    mock_args.agent_cls = None
+    mock_args.max_iterations = None
+
+    # Create a config with specific values from config.toml
+    config_from_toml = OpenHandsConfig()
+    config_from_toml.default_agent = 'ConfigTomlAgent'
+    config_from_toml.max_iterations = 100
+
+    # Load config
+    with patch(
+        'openhands.core.config.utils.load_openhands_config',
+        return_value=config_from_toml,
+    ):
+        config = setup_config_from_args(mock_args)
+
+    # Verify config.toml values are preserved when CLI args are None
+    assert config.default_agent == 'ConfigTomlAgent'
+    assert config.max_iterations == 100

tests/unit/test_get_repository_microagents.py

@@ -13,6 +13,7 @@ from openhands.integrations.service_types import (
     Repository,
 )
 from openhands.microagent.types import MicroagentContentResponse
+from openhands.server.dependencies import check_session_api_key
 from openhands.server.routes.git import app as git_app
 from openhands.server.user_auth import (
     get_access_token,
@@ -49,10 +50,15 @@ def test_client():
     def mock_get_user_id():
         return 'test_user'
 
+    def mock_check_session_api_key():
+        # Mock session API key check to always pass for tests
+        return None
+
     # Override the dependencies in the app
     app.dependency_overrides[get_provider_tokens] = mock_get_provider_tokens
     app.dependency_overrides[get_access_token] = mock_get_access_token
     app.dependency_overrides[get_user_id] = mock_get_user_id
+    app.dependency_overrides[check_session_api_key] = mock_check_session_api_key
 
     yield TestClient(app)
 

tests/unit/test_middleware.py

@@ -46,24 +46,32 @@ def test_localhost_cors_middleware_init_without_env_var():
 
 
 def test_localhost_cors_middleware_is_allowed_origin_localhost(app):
-    """Test that localhost origins are allowed regardless of port."""
-    app.add_middleware(LocalhostCORSMiddleware)
-    client = TestClient(app)
+    """Test that localhost origins are allowed regardless of port when no specific origins are configured."""
+    # Test without setting PERMITTED_CORS_ORIGINS to trigger localhost behavior
+    with patch.dict(os.environ, {}, clear=True):
+        app.add_middleware(LocalhostCORSMiddleware)
+        client = TestClient(app)
 
-    # Test with localhost
-    response = client.get('/test', headers={'Origin': 'http://localhost:8000'})
-    assert response.status_code == 200
-    assert response.headers['access-control-allow-origin'] == 'http://localhost:8000'
+        # Test with localhost
+        response = client.get('/test', headers={'Origin': 'http://localhost:8000'})
+        assert response.status_code == 200
+        assert (
+            response.headers['access-control-allow-origin'] == 'http://localhost:8000'
+        )
 
-    # Test with different port
-    response = client.get('/test', headers={'Origin': 'http://localhost:3000'})
-    assert response.status_code == 200
-    assert response.headers['access-control-allow-origin'] == 'http://localhost:3000'
+        # Test with different port
+        response = client.get('/test', headers={'Origin': 'http://localhost:3000'})
+        assert response.status_code == 200
+        assert (
+            response.headers['access-control-allow-origin'] == 'http://localhost:3000'
+        )
 
-    # Test with 127.0.0.1
-    response = client.get('/test', headers={'Origin': 'http://127.0.0.1:8000'})
-    assert response.status_code == 200
-    assert response.headers['access-control-allow-origin'] == 'http://127.0.0.1:8000'
+        # Test with 127.0.0.1
+        response = client.get('/test', headers={'Origin': 'http://127.0.0.1:8000'})
+        assert response.status_code == 200
+        assert (
+            response.headers['access-control-allow-origin'] == 'http://127.0.0.1:8000'
+        )
 
 
 def test_localhost_cors_middleware_is_allowed_origin_non_localhost(app):
@@ -87,14 +95,15 @@ def test_localhost_cors_middleware_is_allowed_origin_non_localhost(app):
 
 def test_localhost_cors_middleware_missing_origin(app):
     """Test behavior when Origin header is missing."""
-    app.add_middleware(LocalhostCORSMiddleware)
-    client = TestClient(app)
+    with patch.dict(os.environ, {}, clear=True):
+        app.add_middleware(LocalhostCORSMiddleware)
+        client = TestClient(app)
 
-    # Test without Origin header
-    response = client.get('/test')
-    assert response.status_code == 200
-    # There should be no access-control-allow-origin header
-    assert 'access-control-allow-origin' not in response.headers
+        # Test without Origin header
+        response = client.get('/test')
+        assert response.status_code == 200
+        # There should be no access-control-allow-origin header
+        assert 'access-control-allow-origin' not in response.headers
 
 
 def test_localhost_cors_middleware_inheritance():