Add comprehensive tests for conversation metrics API

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

.github/workflows/ghcr-build.yml

@@ -41,8 +41,10 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@v4
+        with:
+          ref: ${{ github.event.pull_request.head.sha }}
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v3.4.0
+        uses: docker/setup-qemu-action@v3.6.0
         with:
           image: tonistiigi/binfmt:latest
       - name: Login to GHCR
@@ -90,8 +92,10 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@v4
+        with:
+          ref: ${{ github.event.pull_request.head.sha }}
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v3.4.0
+        uses: docker/setup-qemu-action@v3.6.0
         with:
           image: tonistiigi/binfmt:latest
       - name: Login to GHCR
@@ -154,6 +158,8 @@ jobs:
         base_image: ['nikolaik']
     steps:
       - uses: actions/checkout@v4
+        with:
+          ref: ${{ github.event.pull_request.head.sha }}
       - name: Cache Poetry dependencies
         uses: actions/cache@v4
         with:

ISSUE_TRIAGE.md

@@ -2,12 +2,13 @@
 These are the procedures and guidelines on how issues are triaged in this repo by the maintainers.
 
 ## General
-* Most issues must be tagged with **enhancement** or **bug**.
-* Issues may be tagged with what it relates to (**backend**, **frontend**, **agent quality**, etc.).
+* All issues must be tagged with **enhancement**, **bug** or **troubleshooting/help**.
+* Issues may be tagged with what it relates to (**agent quality**, **frontend**, **resolver**, etc.).
 
 ## Severity
 * **Low**: Minor issues or affecting single user.
 * **Medium**: Affecting multiple users.
+* **High**: High visibility issues or affecting many users.
 * **Critical**: Affecting all users or potential security issues.
 
 ## Effort
@@ -18,8 +19,14 @@ These are the procedures and guidelines on how issues are triaged in this repo b
 
 ## Not Enough Information
 * User is asked to provide more information (logs, how to reproduce, etc.) when the issue is not clear.
-* If an issue is unclear and the author does not provide more information or respond to a request, the issue may be closed as **not planned** (Usually after a week).
+* If an issue is unclear and the author does not provide more information or respond to a request,
+the issue may be closed as **not planned** (Usually after a week).
 
 ## Multiple Requests/Fixes in One Issue
 * These issues will be narrowed down to one request/fix so the issue is more easily tracked and fixed.
 * Issues may be broken down into multiple issues if required.
+
+## Stale and Auto Closures
+* In order to keep a maintainable backlog, issues that have no activity within 30 days are automatically marked as **Stale**.
+* If issues marked as **Stale** continue to have no activity for 7 more days, they will automatically be closed as not planned.
+* Issues may be reopened by maintainers if deemed important.

config.template.toml

@@ -95,6 +95,11 @@ workspace_base = "./workspace"
 # List of allowed file extensions for uploads
 #file_uploads_allowed_extensions = [".*"]
 
+# Whether to enable the default LLM summarizing condenser when no condenser is specified in config
+# When true, a LLMSummarizingCondenserConfig will be used as the default condenser
+# When false, a NoOpCondenserConfig (no summarization) will be used
+#enable_default_condenser = true
+
 #################################### LLM #####################################
 # Configuration for LLM models (group name starts with 'llm')
 # use 'llm' for the default LLM config
@@ -294,6 +299,69 @@ llm_config = 'gpt3'
 # The security analyzer to use (For Headless / CLI only -  In Web this is overridden by Session Init)
 #security_analyzer = ""
 
+#################################### Condenser #################################
+# Condensers control how conversation history is managed and compressed when
+# the context grows too large. Each agent uses one condenser configuration.
+##############################################################################
+[condenser]
+# The type of condenser to use. Available options:
+# - "noop": No condensing, keeps full history (default)
+# - "observation_masking": Keeps full event structure but masks older observations
+# - "recent": Keeps only recent events and discards older ones
+# - "llm": Uses an LLM to summarize conversation history
+# - "amortized": Intelligently forgets older events while preserving important context
+# - "llm_attention": Uses an LLM to prioritize most relevant context
+type = "noop"
+
+# Examples for each condenser type (uncomment and modify as needed):
+
+# 1. NoOp Condenser - No additional settings needed
+#type = "noop"
+
+# 2. Observation Masking Condenser
+#type = "observation_masking"
+# Number of most-recent events where observations will not be masked
+#attention_window = 100
+
+# 3. Recent Events Condenser
+#type = "recent"
+# Number of initial events to always keep (typically includes task description)
+#keep_first = 1
+# Maximum number of events to keep in history
+#max_events = 100
+
+# 4. LLM Summarizing Condenser
+#type = "llm"
+# Reference to an LLM config to use for summarization
+#llm_config = "condenser"
+# Number of initial events to always keep (typically includes task description)
+#keep_first = 1
+# Maximum size of history before triggering summarization
+#max_size = 100
+
+# 5. Amortized Forgetting Condenser
+#type = "amortized"
+# Number of initial events to always keep (typically includes task description)
+#keep_first = 1
+# Maximum size of history before triggering forgetting
+#max_size = 100
+
+# 6. LLM Attention Condenser
+#type = "llm_attention"
+# Reference to an LLM config to use for attention scoring
+#llm_config = "condenser"
+# Number of initial events to always keep (typically includes task description)
+#keep_first = 1
+# Maximum size of history before triggering attention mechanism
+#max_size = 100
+
+# Example of a custom LLM configuration for condensers that require an LLM
+# If not provided, it falls back to the default LLM
+#[llm.condenser]
+#model = "gpt-4o"
+#temperature = 0.1
+#max_tokens = 1024
+
 #################################### Eval ####################################
 # Configuration for the evaluation, please refer to the specific evaluation
 # plugin for the available options

docs/modules/usage/runtimes.md

@@ -84,3 +84,36 @@ docker run # ...
     -e MODAL_API_TOKEN_ID="your-id" \
     -e MODAL_API_TOKEN_SECRET="your-secret" \
 ```
+
+## Daytona Runtime
+
+Another option is using [Daytona](https://www.daytona.io/) as a runtime provider:
+
+### Step 1: Retrieve Your Daytona API Key
+1. Visit the [Daytona Dashboard](https://app.daytona.io/dashboard/keys).
+2. Click **"Create Key"**.
+3. Enter a name for your key and confirm the creation.
+4. Once the key is generated, copy it.
+
+### Step 2: Set Your API Key as an Environment Variable
+Run the following command in your terminal, replacing `<your-api-key>` with the actual key you copied:
+```bash
+export DAYTONA_API_KEY="<your-api-key>"
+```
+
+This step ensures that OpenHands can authenticate with the Daytona platform when it runs.
+
+### Step 3: Run OpenHands Locally Using Docker
+To start the latest version of OpenHands on your machine, execute the following command in your terminal:
+```bash
+bash -i <(curl -sL https://get.daytona.io/openhands)
+```
+
+#### What This Command Does:
+- Downloads the latest OpenHands release script.
+- Runs the script in an interactive Bash session.
+- Automatically pulls and runs the OpenHands container using Docker.
+
+Once executed, OpenHands should be running locally and ready for use.
+
+For more details and manual initialization, view the entire [README.md](https://github.com/All-Hands-AI/OpenHands/blob/main/openhands/runtime/impl/daytona/README.md)

docs/src/components/Demo/Demo.tsx

@@ -5,9 +5,6 @@ export function Demo() {
   const videoRef = React.useRef<HTMLVideoElement>(null);
 
   return (
-    <div
-      style={{ paddingBottom: "10px", paddingTop: "10px", textAlign: "center" }}
-    >
       <video
         playsInline
         autoPlay={true}
@@ -20,6 +17,5 @@ export function Demo() {
       >
         <source src="img/teaser.mp4" type="video/mp4"></source>
       </video>
-    </div>
   );
 }

docs/src/components/Demo/index.module.css

@@ -1,6 +1,5 @@
 .demo {
   width: 100%;
-  padding: 30px;
   max-width: 800px;
   text-align: center;
   border-radius: 40px;

docs/src/components/HomepageHeader/HomepageHeader.tsx

@@ -17,6 +17,29 @@ export function HomepageHeader() {
 
         <p className="header-subtitle">{siteConfig.tagline}</p>
 
+        <div style={{
+          textAlign: 'center',
+          fontSize: '1.2rem',
+          maxWidth: '800px',
+          margin: '0 auto',
+          padding: '0rem 0rem 1rem'
+        }}>
+        <p style={{ margin: '0' }}>
+          Use AI to tackle the toil in your backlog. Our agents have all the same tools as a human developer: they can modify code, run commands, browse the web,
+          call APIs, and yes-even copy code snippets from StackOverflow.
+          <br/>
+          <Link to="https://docs.all-hands.dev/modules/usage/installation"
+            style={{
+              textDecoration: 'underline',
+              display: 'inline-block',
+              marginTop: '0.5rem'
+            }}
+          >
+            Get started with OpenHands.
+          </Link>
+        </p>
+      </div>
+
         <div align="center" className="header-links">
           <a href="https://github.com/All-Hands-AI/OpenHands/graphs/contributors"><img src="https://img.shields.io/github/contributors/All-Hands-AI/OpenHands?style=for-the-badge&color=blue" alt="Contributors" /></a>
           <a href="https://github.com/All-Hands-AI/OpenHands/stargazers"><img src="https://img.shields.io/github/stars/All-Hands-AI/OpenHands?style=for-the-badge&color=blue" alt="Stargazers" /></a>
@@ -27,12 +50,9 @@ export function HomepageHeader() {
           <a href="https://discord.gg/ESHStjSjD4"><img src="https://img.shields.io/badge/Discord-Join%20Us-purple?logo=discord&logoColor=white&style=for-the-badge" alt="Join our Discord community" /></a>
           <a href="https://github.com/All-Hands-AI/OpenHands/blob/main/CREDITS.md"><img src="https://img.shields.io/badge/Project-Credits-blue?style=for-the-badge&color=FFE165&logo=github&logoColor=white" alt="Credits" /></a>
           <br/>
-          <a href="https://docs.all-hands.dev/modules/usage/getting-started"><img src="https://img.shields.io/badge/Documentation-000?logo=googledocs&logoColor=FFE165&style=for-the-badge" alt="Check out the documentation" /></a>
           <a href="https://arxiv.org/abs/2407.16741"><img src="https://img.shields.io/badge/Paper%20on%20Arxiv-000?logoColor=FFE165&logo=arxiv&style=for-the-badge" alt="Paper on Arxiv" /></a>
           <a href="https://huggingface.co/spaces/OpenHands/evaluation"><img src="https://img.shields.io/badge/Benchmark%20score-000?logoColor=FFE165&logo=huggingface&style=for-the-badge" alt="Evaluation Benchmark Score" /></a>
         </div>
-
-        <Demo />
       </div>
     </div>
   );

docs/src/css/homepageHeader.css

@@ -1,14 +1,14 @@
 /* homepageHeader.css */
 
 .homepage-header {
-  height: 800px;
+  padding: 1rem 0;
 }
 
 .header-content {
   display: flex;
   flex-direction: column;
   align-items: center;
-  padding: 2rem;
+  padding: 1rem;
   font-weight: 300;
   width: 100%;
 }
@@ -25,6 +25,7 @@
 
 .header-subtitle {
   font-size: 1.5rem;
+  margin: 0.5rem 0;
 }
 
 .header-links {

docs/src/pages/index.tsx

@@ -2,15 +2,7 @@ import useDocusaurusContext from '@docusaurus/useDocusaurusContext';
 import Layout from '@theme/Layout';
 import { HomepageHeader } from '../components/HomepageHeader/HomepageHeader';
 import { translate } from '@docusaurus/Translate';
-
-export function Header({ title, summary }): JSX.Element {
-  return (
-    <div>
-      <h1>{title}</h1>
-      <h2 style={{ fontSize: '3rem' }}>{summary}</h2>
-    </div>
-  );
-}
+import { Demo } from "../components/Demo/Demo";
 
 export default function Home(): JSX.Element {
   const { siteConfig } = useDocusaurusContext();
@@ -23,11 +15,14 @@ export default function Home(): JSX.Element {
       })}
     >
       <HomepageHeader />
-      <div style={{ textAlign: 'center', padding: '2rem' }}>
-        <br />
+
+      <div style={{ textAlign: 'center', padding: '1rem 0' }}>
+        <Demo />
+      </div>
+
+      <div style={{ textAlign: 'center', padding: '0.5rem 2rem 1.5rem' }}>
         <h2>Most Popular Links</h2>
         <ul style={{ listStyleType: 'none'}}>
-          <li><a href="/modules/usage/Installation">How to Run OpenHands</a></li>
           <li><a href="/modules/usage/prompting/microagents-repo">Customizing OpenHands to a repository</a></li>
           <li><a href="/modules/usage/how-to/github-action">Integrating OpenHands with Github</a></li>
           <li><a href="/modules/usage/llms#model-recommendations">Recommended models to use</a></li>

evaluation/benchmarks/EDA/run_infer.py

@@ -24,6 +24,7 @@ from openhands.core.config import (
 from openhands.core.logger import openhands_logger as logger
 from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import MessageAction
+from openhands.utils.async_utils import call_async_from_sync
 
 game = None
 
@@ -121,6 +122,7 @@ def process_instance(
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
 
     state: State | None = asyncio.run(
         run_controller(

evaluation/benchmarks/agent_bench/run_infer.py

@@ -34,6 +34,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import AgentFinishAction, CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 
 def get_config(
@@ -210,6 +211,7 @@ def process_instance(
     # =============================================
 
     runtime: Runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
 
     initialize_runtime(runtime, instance=instance)
 

evaluation/benchmarks/aider_bench/run_infer.py

@@ -34,6 +34,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 # Configure visibility of unit tests to the Agent.
 USE_UNIT_TESTS = os.environ.get('USE_UNIT_TESTS', 'false').lower() == 'true'
@@ -203,7 +204,7 @@ def process_instance(
     # =============================================
 
     runtime: Runtime = create_runtime(config)
-
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance=instance)
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state

evaluation/benchmarks/biocoder/run_infer.py

@@ -31,6 +31,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 AGENT_CLS_TO_FAKE_USER_RESPONSE_FN = {
     'CodeActAgent': functools.partial(
@@ -274,6 +275,7 @@ def process_instance(
     instruction += AGENT_CLS_TO_INST_SUFFIX[metadata.agent_class]
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance)
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state

evaluation/benchmarks/bird/run_infer.py

@@ -34,6 +34,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 
 def codeact_user_response(state: State) -> str:
@@ -399,6 +400,7 @@ def process_instance(
     instruction += AGENT_CLS_TO_INST_SUFFIX[metadata.agent_class]
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance)
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state

evaluation/benchmarks/browsing_delegation/run_infer.py

@@ -25,6 +25,7 @@ from openhands.core.config import (
 from openhands.core.logger import openhands_logger as logger
 from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import MessageAction
+from openhands.utils.async_utils import call_async_from_sync
 
 # Only CodeActAgent can delegate to BrowsingAgent
 SUPPORTED_AGENT_CLS = {'CodeActAgent'}
@@ -74,6 +75,7 @@ def process_instance(
     )
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
 
     state: State | None = asyncio.run(
         run_controller(

evaluation/benchmarks/commit0_bench/run_infer.py

@@ -35,6 +35,7 @@ from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation, ErrorObservation
 from openhands.events.serialization.event import event_to_dict
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 from openhands.utils.shutdown_listener import sleep_if_should_continue
 
 USE_HINT_TEXT = os.environ.get('USE_HINT_TEXT', 'false').lower() == 'true'
@@ -394,6 +395,7 @@ def process_instance(
         logger.info(f'Starting evaluation for instance {instance.instance_id}.')
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     try:
         initialize_runtime(runtime, instance)
 

evaluation/benchmarks/discoverybench/run_infer.py

@@ -34,6 +34,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import AgentFinishAction, CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 EVALUATION_LLM = 'gpt-4-1106-preview'
 
@@ -281,6 +282,7 @@ def process_instance(
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance.data_files)
 
     state: State | None = asyncio.run(

evaluation/benchmarks/gaia/run_infer.py

@@ -31,6 +31,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import AgentFinishAction, CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 DATASET_CACHE_DIR = os.path.join(os.path.dirname(__file__), 'data')
 
@@ -148,6 +149,7 @@ def process_instance(
     logger.info(f'Instruction:\n{instruction}', extra={'msg_type': 'OBSERVATION'})
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance)
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state

evaluation/benchmarks/gorilla/ast_eval_hf.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 # This file is modified from https://github.com/ShishirPatil/gorilla/blob/main/eval/eval-scripts/ast_eval_hf.py
 
+import tree_sitter_python as tspython
 from tree_sitter import Language, Parser
 
 
@@ -39,10 +40,9 @@ def get_all_sub_trees(root_node):
 
 
 # Parse the program into AST trees
-def ast_parse(candidate, lang='python'):
-    LANGUAGE = Language('evaluation/gorilla/my-languages.so', lang)
-    parser = Parser()
-    parser.set_language(LANGUAGE)
+def ast_parse(candidate):
+    LANGUAGE = Language(tspython.language())
+    parser = Parser(LANGUAGE)
 
     candidate_tree = parser.parse(bytes(candidate, 'utf8')).root_node
     return candidate_tree

evaluation/benchmarks/gorilla/ast_eval_tf.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 # This file is modified from https://github.com/ShishirPatil/gorilla/blob/main/eval/eval-scripts/ast_eval_tf.py
 
+import tree_sitter_python as tspython
 from tree_sitter import Language, Parser
 
 
@@ -39,10 +40,9 @@ def get_all_sub_trees(root_node):
 
 
 # Parse the program into AST trees
-def ast_parse(candidate, lang='python'):
-    LANGUAGE = Language('evaluation/gorilla/my-languages.so', lang)
-    parser = Parser()
-    parser.set_language(LANGUAGE)
+def ast_parse(candidate):
+    LANGUAGE = Language(tspython.language())
+    parser = Parser(LANGUAGE)
 
     candidate_tree = parser.parse(bytes(candidate, 'utf8')).root_node
     return candidate_tree

evaluation/benchmarks/gorilla/ast_eval_th.py

@@ -13,6 +13,7 @@
 # limitations under the License.
 # This file is modified from https://github.com/ShishirPatil/gorilla/blob/main/eval/eval-scripts/ast_eval_th.py
 
+import tree_sitter_python as tspython
 from tree_sitter import Language, Parser
 
 
@@ -39,10 +40,9 @@ def get_all_sub_trees(root_node):
 
 
 # Parse the program into AST trees
-def ast_parse(candidate, lang='python'):
-    LANGUAGE = Language('evaluation/gorilla/my-languages.so', lang)
-    parser = Parser()
-    parser.set_language(LANGUAGE)
+def ast_parse(candidate):
+    LANGUAGE = Language(tspython.language())
+    parser = Parser(LANGUAGE)
 
     candidate_tree = parser.parse(bytes(candidate, 'utf8')).root_node
     return candidate_tree

evaluation/benchmarks/gorilla/run_infer.py

@@ -26,6 +26,7 @@ from openhands.core.config import (
 from openhands.core.logger import openhands_logger as logger
 from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import MessageAction
+from openhands.utils.async_utils import call_async_from_sync
 
 AGENT_CLS_TO_FAKE_USER_RESPONSE_FN = {
     'CodeActAgent': codeact_user_response,
@@ -82,6 +83,7 @@ def process_instance(
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     state: State | None = asyncio.run(
         run_controller(
             config=config,

evaluation/benchmarks/gorilla/utils.py

@@ -71,19 +71,19 @@ def fetch_data(url, filename):
 
 def get_data_for_hub(hub: str):
     if hub == 'hf':
-        question_data = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/eval/eval-data/questions/huggingface/questions_huggingface_0_shot.jsonl'
-        api_dataset = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/data/api/huggingface_api.jsonl'
-        apibench = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/data/apibench/huggingface_eval.json'
+        question_data = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/eval/eval-data/questions/huggingface/questions_huggingface_0_shot.jsonl'
+        api_dataset = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/data/api/huggingface_api.jsonl'
+        apibench = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/data/apibench/huggingface_eval.json'
         ast_eval = ast_eval_hf
     elif hub == 'torch':
-        question_data = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/eval/eval-data/questions/torchhub/questions_torchhub_0_shot.jsonl'
-        api_dataset = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/data/api/torchhub_api.jsonl'
-        apibench = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/data/apibench/torchhub_eval.json'
+        question_data = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/eval/eval-data/questions/torchhub/questions_torchhub_0_shot.jsonl'
+        api_dataset = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/data/api/torchhub_api.jsonl'
+        apibench = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/data/apibench/torchhub_eval.json'
         ast_eval = ast_eval_th
     elif hub == 'tf':
-        question_data = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/eval/eval-data/questions/tensorflowhub/questions_tensorflowhub_0_shot.jsonl'
-        api_dataset = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/data/api/tensorflowhub_api.jsonl'
-        apibench = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/main/data/apibench/tensorflow_eval.json'
+        question_data = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/eval/eval-data/questions/tensorflowhub/questions_tensorflowhub_0_shot.jsonl'
+        api_dataset = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/data/api/tensorflowhub_api.jsonl'
+        apibench = 'https://raw.githubusercontent.com/ShishirPatil/gorilla/refs/tags/v1.2/data/apibench/tensorflow_eval.json'
         ast_eval = ast_eval_tf
 
     question_data = fetch_data(question_data, 'question_data.jsonl')

evaluation/benchmarks/gpqa/run_infer.py

@@ -49,6 +49,7 @@ from openhands.events.action import (
     MessageAction,
 )
 from openhands.events.observation import Observation
+from openhands.utils.async_utils import call_async_from_sync
 
 ACTION_FORMAT = """
 <<FINAL_ANSWER||
@@ -214,6 +215,7 @@ Ok now its time to start solving the question. Good luck!
 """
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     state: State | None = asyncio.run(
         run_controller(
             config=config,

evaluation/benchmarks/humanevalfix/run_infer.py

@@ -39,6 +39,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 IMPORT_HELPER = {
     'python': [
@@ -232,6 +233,7 @@ def process_instance(
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance)
     state: State | None = asyncio.run(
         run_controller(

evaluation/benchmarks/logic_reasoning/run_infer.py

@@ -31,6 +31,7 @@ from openhands.events.action import (
 )
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 AGENT_CLS_TO_FAKE_USER_RESPONSE_FN = {
     'CodeActAgent': codeact_user_response,
@@ -206,6 +207,7 @@ def process_instance(
     instruction += AGENT_CLS_TO_INST_SUFFIX[metadata.agent_class]
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance)
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state

evaluation/benchmarks/miniwob/run_infer.py

@@ -41,6 +41,7 @@ from openhands.runtime.browser.browser_env import (
     BROWSER_EVAL_GET_GOAL_ACTION,
     BROWSER_EVAL_GET_REWARDS_ACTION,
 )
+from openhands.utils.async_utils import call_async_from_sync
 
 SUPPORTED_AGENT_CLS = {'BrowsingAgent', 'CodeActAgent'}
 
@@ -145,6 +146,7 @@ def process_instance(
         logger.info(f'Starting evaluation for instance {env_id}.')
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     task_str, obs = initialize_runtime(runtime)
 
     task_str += (

evaluation/benchmarks/mint/run_infer.py

@@ -35,6 +35,7 @@ from openhands.events.action import (
 )
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 
 def codeact_user_response_mint(state: State, task: Task, task_config: dict[str, int]):
@@ -184,6 +185,7 @@ def process_instance(
     )
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime)
 
     state: State | None = asyncio.run(

evaluation/benchmarks/ml_bench/run_infer.py

@@ -43,6 +43,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 config = load_app_config()
 
@@ -234,6 +235,7 @@ def process_instance(instance: Any, metadata: EvalMetadata, reset_logger: bool =
     instruction += AGENT_CLS_TO_INST_SUFFIX[metadata.agent_class]
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance)
 
     # Run the agent

evaluation/benchmarks/scienceagentbench/run_infer.py

@@ -29,6 +29,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 AGENT_CLS_TO_FAKE_USER_RESPONSE_FN = {
     'CodeActAgent': codeact_user_response,
@@ -195,6 +196,7 @@ If the program uses some packages that are incompatible, please figure out alter
 """
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime, instance)
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state

evaluation/benchmarks/swe_bench/run_infer.py

@@ -40,6 +40,7 @@ from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation, ErrorObservation
 from openhands.events.serialization.event import event_to_dict
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 from openhands.utils.shutdown_listener import sleep_if_should_continue
 
 USE_HINT_TEXT = os.environ.get('USE_HINT_TEXT', 'false').lower() == 'true'
@@ -464,6 +465,7 @@ def process_instance(
             f'This is the {runtime_failure_count + 1}th attempt for instance {instance.instance_id}, setting resource factor to {config.sandbox.remote_runtime_resource_factor}'
         )
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
 
     try:
         initialize_runtime(runtime, instance)

evaluation/benchmarks/the_agent_company/browsing.py

@@ -7,7 +7,7 @@ import os
 import re
 from dataclasses import dataclass
 from enum import Enum, auto
-from typing import Dict, List, Optional, Union
+from typing import Dict, List, Union
 
 from openhands.core.logger import openhands_logger as logger
 from openhands.events.action import BrowseInteractiveAction
@@ -133,7 +133,7 @@ def parse_content_to_elements(content: str) -> Dict[str, str]:
     return elements
 
 
-def find_matching_anchor(content: str, selector: str) -> Optional[str]:
+def find_matching_anchor(content: str, selector: str) -> str | None:
     """Find the anchor ID that matches the given selector description"""
     elements = parse_content_to_elements(content)
 

evaluation/benchmarks/the_agent_company/run_infer.py

@@ -28,6 +28,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import BrowserOutputObservation, CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 
 def get_config(
@@ -275,7 +276,7 @@ if __name__ == '__main__':
         args.task_image_name, task_short_name, temp_dir, agent_llm_config, agent_config
     )
     runtime: Runtime = create_runtime(config)
-
+    call_async_from_sync(runtime.connect)
     init_task_env(runtime, args.server_hostname, env_llm_config)
 
     dependencies = load_dependencies(runtime)

evaluation/benchmarks/toolqa/run_infer.py

@@ -27,6 +27,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
 from openhands.events.observation import CmdOutputObservation
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 AGENT_CLS_TO_FAKE_USER_RESPONSE_FN = {
     'CodeActAgent': codeact_user_response,
@@ -104,6 +105,7 @@ def process_instance(instance: Any, metadata: EvalMetadata, reset_logger: bool =
     logger.info(f'Instruction:\n{instruction}', extra={'msg_type': 'OBSERVATION'})
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     initialize_runtime(runtime)
 
     # Here's how you can run the agent (similar to the `main` function) and get the final task state

evaluation/benchmarks/visualwebarena/run_infer.py

@@ -37,6 +37,7 @@ from openhands.runtime.browser.browser_env import (
     BROWSER_EVAL_GET_GOAL_ACTION,
     BROWSER_EVAL_GET_REWARDS_ACTION,
 )
+from openhands.utils.async_utils import call_async_from_sync
 
 SUPPORTED_AGENT_CLS = {'VisualBrowsingAgent'}
 AGENT_CLS_TO_FAKE_USER_RESPONSE_FN = {
@@ -159,6 +160,8 @@ def process_instance(
         logger.info(f'Starting evaluation for instance {env_id}.')
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
+
     task_str, goal_image_urls = initialize_runtime(runtime)
     initial_user_action = MessageAction(content=task_str, image_urls=goal_image_urls)
     state: State | None = asyncio.run(

evaluation/benchmarks/webarena/run_infer.py

@@ -36,6 +36,7 @@ from openhands.runtime.browser.browser_env import (
     BROWSER_EVAL_GET_GOAL_ACTION,
     BROWSER_EVAL_GET_REWARDS_ACTION,
 )
+from openhands.utils.async_utils import call_async_from_sync
 
 SUPPORTED_AGENT_CLS = {'BrowsingAgent'}
 
@@ -144,6 +145,7 @@ def process_instance(
         logger.info(f'Starting evaluation for instance {env_id}.')
 
     runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     task_str = initialize_runtime(runtime)
 
     state: State | None = asyncio.run(

evaluation/integration_tests/run_infer.py

@@ -30,6 +30,7 @@ from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import MessageAction
 from openhands.events.serialization.event import event_to_dict
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 FAKE_RESPONSES = {
     'CodeActAgent': fake_user_response,
@@ -108,6 +109,7 @@ def process_instance(
     # create sandbox and run the agent
     # =============================================
     runtime: Runtime = create_runtime(config)
+    call_async_from_sync(runtime.connect)
     try:
         test_class.initialize_runtime(runtime)
 

frontend/__tests__/components/shared/modals/settings/runtime-size-selector.test.tsx

@@ -1,35 +0,0 @@
-import { screen } from "@testing-library/react";
-import { describe, it, expect } from "vitest";
-import { renderWithProviders } from "test-utils";
-import { RuntimeSizeSelector } from "#/components/shared/modals/settings/runtime-size-selector";
-
-const renderRuntimeSizeSelector = () =>
-  renderWithProviders(<RuntimeSizeSelector isDisabled={false} />);
-
-describe("RuntimeSizeSelector", () => {
-  it("should show both runtime size options", () => {
-    renderRuntimeSizeSelector();
-    // The options are in the hidden select element
-    const select = screen.getByRole("combobox", { hidden: true });
-    expect(select).toHaveValue("1");
-    expect(select).toHaveDisplayValue("1x (2 core, 8G)");
-    expect(select.children).toHaveLength(3); // Empty option + 2 size options
-  });
-
-  it("should show the full description text for disabled options", async () => {
-    renderRuntimeSizeSelector();
-
-    // Click the button to open the dropdown
-    const button = screen.getByRole("button", {
-      name: "1x (2 core, 8G) SETTINGS_FORM$RUNTIME_SIZE_LABEL",
-    });
-    button.click();
-
-    // Wait for the dropdown to open and find the description text
-    const description = await screen.findByText(
-      "Runtime sizes over 1 are disabled by default, please contact contact@all-hands.dev to get access to larger runtimes.",
-    );
-    expect(description).toBeInTheDocument();
-    expect(description).toHaveClass("whitespace-normal", "break-words");
-  });
-});

frontend/package.json

@@ -7,47 +7,47 @@
     "node": ">=20.0.0"
   },
   "dependencies": {
-    "@heroui/react": "2.6.14",
+    "@heroui/react": "2.7.4",
     "@monaco-editor/react": "^4.7.0-rc.0",
-    "@react-router/node": "^7.1.5",
-    "@react-router/serve": "^7.1.5",
+    "@react-router/node": "^7.2.0",
+    "@react-router/serve": "^7.2.0",
     "@react-types/shared": "^3.27.0",
-    "@reduxjs/toolkit": "^2.5.1",
+    "@reduxjs/toolkit": "^2.6.0",
     "@stripe/react-stripe-js": "^3.1.1",
-    "@stripe/stripe-js": "^5.5.0",
-    "@tanstack/react-query": "^5.66.7",
+    "@stripe/stripe-js": "^5.7.0",
+    "@tanstack/react-query": "^5.66.11",
     "@vitejs/plugin-react": "^4.3.2",
     "@xterm/addon-fit": "^0.10.0",
     "@xterm/xterm": "^5.4.0",
-    "axios": "^1.7.9",
+    "axios": "^1.8.1",
     "clsx": "^2.1.1",
     "eslint-config-airbnb-typescript": "^18.0.0",
-    "framer-motion": "^12.4.4",
+    "framer-motion": "^12.4.7",
     "i18next": "^24.2.2",
-    "i18next-browser-languagedetector": "^8.0.3",
+    "i18next-browser-languagedetector": "^8.0.4",
     "i18next-http-backend": "^3.0.2",
-    "isbot": "^5.1.22",
-    "jose": "^5.10.0",
+    "isbot": "^5.1.23",
+    "jose": "^6.0.8",
     "monaco-editor": "^0.52.2",
-    "posthog-js": "^1.219.3",
+    "posthog-js": "^1.225.1",
     "react": "^19.0.0",
     "react-dom": "^19.0.0",
     "react-highlight": "^0.15.0",
     "react-hot-toast": "^2.5.1",
     "react-i18next": "^15.4.1",
-    "react-icons": "^5.4.0",
-    "react-markdown": "^9.0.3",
+    "react-icons": "^5.5.0",
+    "react-markdown": "^10.0.1",
     "react-redux": "^9.2.0",
-    "react-router": "^7.1.5",
+    "react-router": "^7.2.0",
     "react-syntax-highlighter": "^15.6.1",
     "react-textarea-autosize": "^8.5.7",
     "remark-gfm": "^4.0.1",
     "sirv-cli": "^3.0.1",
     "socket.io-client": "^4.8.1",
-    "tailwind-merge": "^3.0.1",
-    "vite": "^6.1.0",
+    "tailwind-merge": "^3.0.2",
+    "vite": "^6.2.0",
     "web-vitals": "^3.5.2",
-    "ws": "^8.18.0"
+    "ws": "^8.18.1"
   },
   "scripts": {
     "dev": "npm run make-i18n && cross-env VITE_MOCK_API=false react-router dev",
@@ -81,14 +81,14 @@
   "devDependencies": {
     "@mswjs/socket.io-binding": "^0.1.1",
     "@playwright/test": "^1.50.1",
-    "@react-router/dev": "^7.1.5",
+    "@react-router/dev": "^7.2.0",
     "@tailwindcss/typography": "^0.5.16",
     "@tanstack/eslint-plugin-query": "^5.66.1",
     "@testing-library/dom": "^10.4.0",
     "@testing-library/jest-dom": "^6.6.1",
     "@testing-library/react": "^16.2.0",
     "@testing-library/user-event": "^14.6.1",
-    "@types/node": "^22.13.4",
+    "@types/node": "^22.13.8",
     "@types/react": "^19.0.8",
     "@types/react-dom": "^19.0.3",
     "@types/react-highlight": "^0.12.8",
@@ -96,13 +96,13 @@
     "@types/ws": "^8.5.14",
     "@typescript-eslint/eslint-plugin": "^7.18.0",
     "@typescript-eslint/parser": "^7.18.0",
-    "@vitest/coverage-v8": "^3.0.6",
+    "@vitest/coverage-v8": "^3.0.7",
     "autoprefixer": "^10.4.20",
     "cross-env": "^7.0.3",
     "eslint": "^8.57.0",
     "eslint-config-airbnb": "^19.0.4",
     "eslint-config-airbnb-typescript": "^18.0.0",
-    "eslint-config-prettier": "^10.0.1",
+    "eslint-config-prettier": "^10.0.2",
     "eslint-plugin-import": "^2.29.1",
     "eslint-plugin-jsx-a11y": "^6.10.2",
     "eslint-plugin-prettier": "^5.2.3",
@@ -113,10 +113,10 @@
     "lint-staged": "^15.4.3",
     "msw": "^2.6.6",
     "postcss": "^8.5.2",
-    "prettier": "^3.5.1",
-    "stripe": "^17.5.0",
+    "prettier": "^3.5.3",
+    "stripe": "^17.7.0",
     "tailwindcss": "^3.4.17",
-    "typescript": "^5.7.3",
+    "typescript": "^5.8.2",
     "vite-plugin-svgr": "^4.2.0",
     "vite-tsconfig-paths": "^5.1.4",
     "vitest": "^3.0.2"

frontend/src/components/features/chat/chat-message.tsx

@@ -57,18 +57,19 @@ export function ChatMessage({
         onClick={handleCopyToClipboard}
         mode={isCopy ? "copied" : "copy"}
       />
-      <Markdown
-        className="text-sm overflow-auto break-words"
-        components={{
-          code,
-          ul,
-          ol,
-          a: anchor,
-        }}
-        remarkPlugins={[remarkGfm]}
-      >
-        {message}
-      </Markdown>
+      <div className="text-sm overflow-auto break-words">
+        <Markdown
+          components={{
+            code,
+            ul,
+            ol,
+            a: anchor,
+          }}
+          remarkPlugins={[remarkGfm]}
+        >
+          {message}
+        </Markdown>
+      </div>
       {children}
     </article>
   );

frontend/src/components/features/chat/expandable-message.tsx

@@ -123,17 +123,18 @@ export function ExpandableMessage({
           )}
         </div>
         {(!headline || showDetails) && (
-          <Markdown
-            className="text-sm overflow-auto"
-            components={{
-              code,
-              ul,
-              ol,
-            }}
-            remarkPlugins={[remarkGfm]}
-          >
-            {details}
-          </Markdown>
+          <div className="text-sm overflow-auto">
+            <Markdown
+              components={{
+                code,
+                ul,
+                ol,
+              }}
+              remarkPlugins={[remarkGfm]}
+            >
+              {details}
+            </Markdown>
+          </div>
         )}
       </div>
     </div>

frontend/src/components/features/github/github-repo-selector.tsx

@@ -99,7 +99,6 @@ export function GitHubRepositorySelector({
             <AutocompleteItem
               data-testid="github-repo-item"
               key={repo.id}
-              value={repo.id}
               className="data-[selected=true]:bg-default-100"
               textValue={repo.full_name}
             >
@@ -114,7 +113,6 @@ export function GitHubRepositorySelector({
             <AutocompleteItem
               data-testid="github-repo-item"
               key={repo.id}
-              value={repo.id}
               className="data-[selected=true]:bg-default-100"
               textValue={repo.full_name}
             >

frontend/src/components/features/settings/settings-dropdown-input.tsx

@@ -1,9 +1,10 @@
 import { Autocomplete, AutocompleteItem } from "@heroui/react";
+import { ReactNode } from "react";
 import { OptionalTag } from "./optional-tag";
 
 interface SettingsDropdownInputProps {
   testId: string;
-  label: string;
+  label: ReactNode;
   name: string;
   items: { key: React.Key; label: string }[];
   showOptionalTag?: boolean;
@@ -29,7 +30,7 @@ export function SettingsDropdownInput({
         {showOptionalTag && <OptionalTag />}
       </div>
       <Autocomplete
-        aria-label={label}
+        aria-label={typeof label === "string" ? label : name}
         data-testid={testId}
         name={name}
         defaultItems={items}

frontend/src/components/shared/form-fieldset.tsx

@@ -1,43 +0,0 @@
-import { Autocomplete, AutocompleteItem } from "@heroui/react";
-
-interface FormFieldsetProps {
-  id: string;
-  label: string;
-  items: { key: string; value: string }[];
-  defaultSelectedKey?: string;
-  isClearable?: boolean;
-}
-
-export function FormFieldset({
-  id,
-  label,
-  items,
-  defaultSelectedKey,
-  isClearable,
-}: FormFieldsetProps) {
-  return (
-    <fieldset className="flex flex-col gap-2">
-      <label htmlFor={id} className="font-[500] text-[#A3A3A3] text-xs">
-        {label}
-      </label>
-      <Autocomplete
-        id={id}
-        name={id}
-        aria-label={label}
-        defaultSelectedKey={defaultSelectedKey}
-        isClearable={isClearable}
-        inputProps={{
-          classNames: {
-            inputWrapper: "bg-[#27272A] rounded-md text-sm px-3 py-[10px]",
-          },
-        }}
-      >
-        {items.map((item) => (
-          <AutocompleteItem key={item.key} value={item.key}>
-            {item.value}
-          </AutocompleteItem>
-        ))}
-      </Autocomplete>
-    </fieldset>
-  );
-}

frontend/src/components/shared/inputs/agent-input.tsx

@@ -1,46 +0,0 @@
-import { Autocomplete, AutocompleteItem } from "@heroui/react";
-import { useTranslation } from "react-i18next";
-import { I18nKey } from "#/i18n/declaration";
-
-interface AgentInputProps {
-  isDisabled: boolean;
-  defaultValue: string;
-  agents: string[];
-}
-
-export function AgentInput({
-  isDisabled,
-  defaultValue,
-  agents,
-}: AgentInputProps) {
-  const { t } = useTranslation();
-
-  return (
-    <fieldset data-testid="agent-selector" className="flex flex-col gap-2">
-      <label htmlFor="agent" className="font-[500] text-[#A3A3A3] text-xs">
-        {t(I18nKey.SETTINGS_FORM$AGENT_LABEL)}
-      </label>
-      <Autocomplete
-        isDisabled={isDisabled}
-        isRequired
-        id="agent"
-        aria-label="Agent"
-        data-testid="agent-input"
-        name="agent"
-        defaultSelectedKey={defaultValue}
-        isClearable={false}
-        inputProps={{
-          classNames: {
-            inputWrapper: "bg-[#27272A] rounded-md text-sm px-3 py-[10px]",
-          },
-        }}
-      >
-        {agents.map((agent) => (
-          <AutocompleteItem key={agent} value={agent}>
-            {agent}
-          </AutocompleteItem>
-        ))}
-      </Autocomplete>
-    </fieldset>
-  );
-}

frontend/src/components/shared/inputs/security-analyzers-input.tsx

@@ -1,46 +0,0 @@
-import { Autocomplete, AutocompleteItem } from "@heroui/react";
-import { useTranslation } from "react-i18next";
-import { I18nKey } from "#/i18n/declaration";
-
-interface SecurityAnalyzerInputProps {
-  isDisabled: boolean;
-  defaultValue: string;
-  securityAnalyzers: string[];
-}
-
-export function SecurityAnalyzerInput({
-  isDisabled,
-  defaultValue,
-  securityAnalyzers,
-}: SecurityAnalyzerInputProps) {
-  const { t } = useTranslation();
-
-  return (
-    <fieldset className="flex flex-col gap-2">
-      <label
-        htmlFor="security-analyzer"
-        className="font-[500] text-[#A3A3A3] text-xs"
-      >
-        {t(I18nKey.SETTINGS_FORM$SECURITY_ANALYZER_LABEL)}
-      </label>
-      <Autocomplete
-        isDisabled={isDisabled}
-        id="security-analyzer"
-        name="security-analyzer"
-        aria-label="Security Analyzer"
-        defaultSelectedKey={defaultValue}
-        inputProps={{
-          classNames: {
-            inputWrapper: "bg-[#27272A] rounded-md text-sm px-3 py-[10px]",
-          },
-        }}
-      >
-        {securityAnalyzers.map((analyzer) => (
-          <AutocompleteItem key={analyzer} value={analyzer}>
-            {analyzer}
-          </AutocompleteItem>
-        ))}
-      </Autocomplete>
-    </fieldset>
-  );
-}

frontend/src/components/shared/modals/settings/model-selector.tsx

@@ -100,7 +100,6 @@ export function ModelSelector({
                 <AutocompleteItem
                   data-testid={`provider-item-${provider}`}
                   key={provider}
-                  value={provider}
                 >
                   {mapProvider(provider)}
                 </AutocompleteItem>
@@ -110,7 +109,7 @@ export function ModelSelector({
             {Object.keys(models)
               .filter((provider) => !VERIFIED_PROVIDERS.includes(provider))
               .map((provider) => (
-                <AutocompleteItem key={provider} value={provider}>
+                <AutocompleteItem key={provider}>
                   {mapProvider(provider)}
                 </AutocompleteItem>
               ))}
@@ -148,9 +147,7 @@ export function ModelSelector({
             {models[selectedProvider || ""]?.models
               .filter((model) => VERIFIED_MODELS.includes(model))
               .map((model) => (
-                <AutocompleteItem key={model} value={model}>
-                  {model}
-                </AutocompleteItem>
+                <AutocompleteItem key={model}>{model}</AutocompleteItem>
               ))}
           </AutocompleteSection>
           <AutocompleteSection title="Others">
@@ -160,7 +157,6 @@ export function ModelSelector({
                 <AutocompleteItem
                   data-testid={`model-item-${model}`}
                   key={model}
-                  value={model}
                 >
                   {model}
                 </AutocompleteItem>

frontend/src/components/shared/modals/settings/runtime-size-selector.tsx

@@ -1,57 +0,0 @@
-import { useTranslation } from "react-i18next";
-import { Select, SelectItem } from "@heroui/react";
-import { I18nKey } from "#/i18n/declaration";
-
-interface RuntimeSizeSelectorProps {
-  isDisabled: boolean;
-  defaultValue?: number;
-}
-
-export function RuntimeSizeSelector({
-  isDisabled,
-  defaultValue,
-}: RuntimeSizeSelectorProps) {
-  const { t } = useTranslation();
-
-  return (
-    <fieldset className="flex flex-col gap-2">
-      <label
-        htmlFor="runtime-size"
-        className="font-[500] text-[#A3A3A3] text-xs"
-      >
-        {t(I18nKey.SETTINGS_FORM$RUNTIME_SIZE_LABEL)}
-      </label>
-      <Select
-        data-testid="runtime-size"
-        id="runtime-size"
-        name="runtime-size"
-        defaultSelectedKeys={[String(defaultValue || 1)]}
-        selectedKeys={[String(defaultValue || 1)]}
-        isDisabled={isDisabled}
-        selectionMode="single"
-        disallowEmptySelection
-        aria-label={t(I18nKey.SETTINGS_FORM$RUNTIME_SIZE_LABEL)}
-        classNames={{
-          trigger: "bg-[#27272A] rounded-md text-sm px-3 py-[10px]",
-        }}
-      >
-        <SelectItem key="1" value={1}>
-          1x (2 core, 8G)
-        </SelectItem>
-        <SelectItem
-          key="2"
-          value={2}
-          isDisabled
-          classNames={{
-            description:
-              "whitespace-normal break-words min-w-[300px] max-w-[300px]",
-            base: "min-w-[300px] max-w-[300px]",
-          }}
-          description="Runtime sizes over 1 are disabled by default, please contact contact@all-hands.dev to get access to larger runtimes."
-        >
-          2x (4 core, 16G)
-        </SelectItem>
-      </Select>
-    </fieldset>
-  );
-}

frontend/src/components/shared/modals/settings/settings-form.tsx

@@ -43,15 +43,18 @@ export function SettingsForm({ settings, models, onClose }: SettingsFormProps) {
   const handleFormSubmission = async (formData: FormData) => {
     const newSettings = extractSettings(formData);
 
-    await saveUserSettings(newSettings);
-    onClose();
-    resetOngoingSession();
+    await saveUserSettings(newSettings, {
+      onSuccess: () => {
+        onClose();
+        resetOngoingSession();
 
-    posthog.capture("settings_saved", {
-      LLM_MODEL: newSettings.LLM_MODEL,
-      LLM_API_KEY: newSettings.LLM_API_KEY ? "SET" : "UNSET",
-      REMOTE_RUNTIME_RESOURCE_FACTOR:
-        newSettings.REMOTE_RUNTIME_RESOURCE_FACTOR,
+        posthog.capture("settings_saved", {
+          LLM_MODEL: newSettings.LLM_MODEL,
+          LLM_API_KEY: newSettings.LLM_API_KEY ? "SET" : "UNSET",
+          REMOTE_RUNTIME_RESOURCE_FACTOR:
+            newSettings.REMOTE_RUNTIME_RESOURCE_FACTOR,
+        });
+      },
     });
   };
 

frontend/src/hooks/query/use-settings.ts

@@ -3,7 +3,6 @@ import React from "react";
 import posthog from "posthog-js";
 import OpenHands from "#/api/open-hands";
 import { useAuth } from "#/context/auth-context";
-import { useConfig } from "#/hooks/query/use-config";
 import { DEFAULT_SETTINGS } from "#/services/settings";
 
 const getSettingsQueryFn = async () => {
@@ -27,12 +26,10 @@ const getSettingsQueryFn = async () => {
 
 export const useSettings = () => {
   const { setGitHubTokenIsSet, githubTokenIsSet } = useAuth();
-  const { data: config } = useConfig();
 
   const query = useQuery({
     queryKey: ["settings", githubTokenIsSet],
     queryFn: getSettingsQueryFn,
-    enabled: config?.APP_MODE !== "saas" || githubTokenIsSet,
     // Only retry if the error is not a 404 because we
     // would want to show the modal immediately if the
     // settings are not found

frontend/src/routes/account-settings.tsx

@@ -278,7 +278,15 @@ function AccountSettings() {
               <SettingsDropdownInput
                 testId="runtime-settings-input"
                 name="runtime-settings-input"
-                label="Runtime Settings"
+                label={
+                  <>
+                    Runtime Settings (
+                    <a href="mailto:contact@all-hands.dev">
+                      get in touch for access
+                    </a>
+                    )
+                  </>
+                }
                 items={REMOTE_RUNTIME_OPTIONS}
                 defaultSelectedKey={settings.REMOTE_RUNTIME_RESOURCE_FACTOR?.toString()}
                 isDisabled

frontend/src/services/observations.ts

@@ -95,6 +95,7 @@ export function handleObservationMessage(message: ObservationMessage) {
             observation,
             extras: {
               path: String(message.extras.path || ""),
+              impl_source: String(message.extras.impl_source || ""),
             },
           }),
         );
@@ -107,6 +108,7 @@ export function handleObservationMessage(message: ObservationMessage) {
             extras: {
               path: String(message.extras.path || ""),
               diff: String(message.extras.diff || ""),
+              impl_source: String(message.extras.impl_source || ""),
             },
           }),
         );

frontend/src/state/chat-slice.ts

@@ -159,9 +159,16 @@ export const chatSlice = createSlice({
           .includes("error:");
       } else if (observationID === "read" || observationID === "edit") {
         // For read/edit operations, we consider it successful if there's content and no error
-        causeMessage.success =
-          observation.payload.content.length > 0 &&
-          !observation.payload.content.toLowerCase().includes("error:");
+
+        if (observation.payload.extras.impl_source === "oh_aci") {
+          causeMessage.success =
+            observation.payload.content.length > 0 &&
+            !observation.payload.content.startsWith("ERROR:\n");
+        } else {
+          causeMessage.success =
+            observation.payload.content.length > 0 &&
+            !observation.payload.content.toLowerCase().includes("error:");
+        }
       }
 
       if (observationID === "run" || observationID === "run_ipython") {

frontend/src/types/core/observations.ts

@@ -63,6 +63,7 @@ export interface ReadObservation extends OpenHandsObservationEvent<"read"> {
   source: "agent";
   extras: {
     path: string;
+    impl_source: string;
   };
 }
 
@@ -71,6 +72,7 @@ export interface EditObservation extends OpenHandsObservationEvent<"edit"> {
   extras: {
     path: string;
     diff: string;
+    impl_source: string;
   };
 }
 

frontend/src/utils/generate-github-auth-url.ts

@@ -6,12 +6,10 @@
  */
 export const generateGitHubAuthUrl = (clientId: string, requestUrl: URL) => {
   const redirectUri = `${requestUrl.origin}/oauth/keycloak/callback`;
-  const baseUrl = `${requestUrl.origin}`
-    .replace("https://", "")
-    .replace("http://", "");
-  const authUrl = baseUrl
+  const authUrl = requestUrl.hostname
     .replace(/(^|\.)staging\.all-hands\.dev$/, "$1auth.staging.all-hands.dev")
-    .replace(/(^|\.)app\.all-hands\.dev$/, "auth.app.all-hands.dev");
-  const scope = "openid email profile";
+    .replace(/(^|\.)app\.all-hands\.dev$/, "auth.app.all-hands.dev")
+    .replace(/(^|\.)localhost$/, "auth.staging.all-hands.dev");
+  const scope = "openid email profile offline_access";
   return `https://${authUrl}/realms/allhands/protocol/openid-connect/auth?client_id=github&response_type=code&redirect_uri=${encodeURIComponent(redirectUri)}&scope=${encodeURIComponent(scope)}`;
 };

microagents/knowledge/docker.md

@@ -53,4 +53,3 @@ To verify Docker is working correctly, run the hello-world container:
 ```bash
 sudo docker run hello-world
 ```
-

openhands/agenthub/codeact_agent/codeact_agent.py

@@ -91,7 +91,7 @@ class CodeActAgent(Agent):
         self.conversation_memory = ConversationMemory(self.prompt_manager)
 
         self.condenser = Condenser.from_config(self.config.condenser)
-        logger.debug(f'Using condenser: {self.condenser}')
+        logger.debug(f'Using condenser: {type(self.condenser)}')
 
     def reset(self) -> None:
         """Resets the CodeAct Agent."""

openhands/controller/agent_controller.py

@@ -248,8 +248,9 @@ class AgentController:
             )
             reported = RuntimeError(
                 'There was an unexpected error while running the agent. Please '
-                f'report this error to the developers. Your session ID is {self.id}. '
-                f'Error type: {e.__class__.__name__}'
+                'report this error to the developers by opening an issue at '
+                'https://github.com/All-Hands-AI/OpenHands. Your session ID is '
+                f' {self.id}. Error type: {e.__class__.__name__}'
             )
             if (
                 isinstance(e, litellm.AuthenticationError)

openhands/core/cli.py

@@ -14,7 +14,12 @@ from openhands.core.config import (
 from openhands.core.logger import openhands_logger as logger
 from openhands.core.loop import run_agent_until_done
 from openhands.core.schema import AgentState
-from openhands.core.setup import create_agent, create_controller, create_runtime
+from openhands.core.setup import (
+    create_agent,
+    create_controller,
+    create_runtime,
+    initialize_repository_for_runtime,
+)
 from openhands.events import EventSource, EventStreamSubscriber
 from openhands.events.action import (
     Action,
@@ -109,7 +114,6 @@ async def main(loop: asyncio.AbstractEventLoop):
         sid=sid,
         headless_mode=True,
         agent=agent,
-        selected_repository=config.sandbox.selected_repo,
     )
 
     controller, _ = create_controller(agent, runtime, config)
@@ -165,6 +169,14 @@ async def main(loop: asyncio.AbstractEventLoop):
 
     await runtime.connect()
 
+    # Initialize repository if needed
+    if config.sandbox.selected_repo:
+        initialize_repository_for_runtime(
+            runtime,
+            agent=agent,
+            selected_repository=config.sandbox.selected_repo,
+        )
+
     if initial_user_action:
         # If there's an initial user action, enqueue it and do not prompt again
         event_stream.add_event(initial_user_action, EventSource.USER)

openhands/core/config/app_config.py

@@ -82,6 +82,7 @@ class AppConfig(BaseModel):
     daytona_target: str = Field(default='us')
     cli_multiline_input: bool = Field(default=False)
     conversation_max_age_seconds: int = Field(default=864000)  # 10 days in seconds
+    enable_default_condenser: bool = Field(default=True)
 
     defaults_dict: ClassVar[dict] = {}
 

openhands/core/config/condenser_config.py

@@ -1,7 +1,8 @@
-from typing import Literal
+from typing import Literal, cast
 
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, Field, ValidationError
 
+from openhands.core import logger
 from openhands.core.config.llm_config import LLMConfig
 
 
@@ -10,17 +11,21 @@ class NoOpCondenserConfig(BaseModel):
 
     type: Literal['noop'] = Field('noop')
 
+    model_config = {'extra': 'forbid'}
+
 
 class ObservationMaskingCondenserConfig(BaseModel):
     """Configuration for ObservationMaskingCondenser."""
 
     type: Literal['observation_masking'] = Field('observation_masking')
     attention_window: int = Field(
-        default=10,
+        default=100,
         description='The number of most-recent events where observations will not be masked.',
         ge=1,
     )
 
+    model_config = {'extra': 'forbid'}
+
 
 class RecentEventsCondenserConfig(BaseModel):
     """Configuration for RecentEventsCondenser."""
@@ -34,9 +39,11 @@ class RecentEventsCondenserConfig(BaseModel):
         ge=0,
     )
     max_events: int = Field(
-        default=10, description='Maximum number of events to keep.', ge=1
+        default=100, description='Maximum number of events to keep.', ge=1
     )
 
+    model_config = {'extra': 'forbid'}
+
 
 class LLMSummarizingCondenserConfig(BaseModel):
     """Configuration for LLMCondenser."""
@@ -49,13 +56,17 @@ class LLMSummarizingCondenserConfig(BaseModel):
     # at least one event by default, because the best guess is that it's the user task
     keep_first: int = Field(
         default=1,
-        description='The number of initial events to condense.',
+        description='Number of initial events to always keep in history.',
         ge=0,
     )
     max_size: int = Field(
-        default=10, description='Maximum number of events to keep.', ge=1
+        default=100,
+        description='Maximum size of the condensed history before triggering forgetting.',
+        ge=2,
     )
 
+    model_config = {'extra': 'forbid'}
+
 
 class AmortizedForgettingCondenserConfig(BaseModel):
     """Configuration for AmortizedForgettingCondenser."""
@@ -74,6 +85,8 @@ class AmortizedForgettingCondenserConfig(BaseModel):
         ge=0,
     )
 
+    model_config = {'extra': 'forbid'}
+
 
 class LLMAttentionCondenserConfig(BaseModel):
     """Configuration for LLMAttentionCondenser."""
@@ -95,7 +108,10 @@ class LLMAttentionCondenserConfig(BaseModel):
         ge=0,
     )
 
+    model_config = {'extra': 'forbid'}
 
+
+# Type alias for convenience
 CondenserConfig = (
     NoOpCondenserConfig
     | ObservationMaskingCondenserConfig
@@ -104,3 +120,121 @@ CondenserConfig = (
     | AmortizedForgettingCondenserConfig
     | LLMAttentionCondenserConfig
 )
+
+
+def condenser_config_from_toml_section(
+    data: dict, llm_configs: dict | None = None
+) -> dict[str, CondenserConfig]:
+    """
+    Create a CondenserConfig instance from a toml dictionary representing the [condenser] section.
+
+    For CondenserConfig, the handling is different since it's a union type. The type of condenser
+    is determined by the 'type' field in the section.
+
+    Example:
+    Parse condenser config like:
+        [condenser]
+        type = "noop"
+
+    For condensers that require an LLM config, you can specify the name of an LLM config:
+        [condenser]
+        type = "llm"
+        llm_config = "my_llm"  # References [llm.my_llm] section
+
+    Args:
+        data: The TOML dictionary representing the [condenser] section.
+        llm_configs: Optional dictionary of LLMConfig objects keyed by name.
+
+    Returns:
+        dict[str, CondenserConfig]: A mapping where the key "condenser" corresponds to the configuration.
+    """
+
+    # Initialize the result mapping
+    condenser_mapping: dict[str, CondenserConfig] = {}
+
+    # Process config
+    try:
+        # Determine which condenser type to use based on 'type' field
+        condenser_type = data.get('type', 'noop')
+
+        # Handle LLM config reference if needed
+        if (
+            condenser_type in ('llm', 'llm_attention')
+            and 'llm_config' in data
+            and isinstance(data['llm_config'], str)
+        ):
+            llm_config_name = data['llm_config']
+            if llm_configs and llm_config_name in llm_configs:
+                # Replace the string reference with the actual LLMConfig object
+                data_copy = data.copy()
+                data_copy['llm_config'] = llm_configs[llm_config_name]
+                config = create_condenser_config(condenser_type, data_copy)
+            else:
+                logger.openhands_logger.warning(
+                    f"LLM config '{llm_config_name}' not found for condenser. Using default LLMConfig."
+                )
+                # Create a default LLMConfig if the referenced one doesn't exist
+                data_copy = data.copy()
+                # Try to use the fallback 'llm' config
+                if llm_configs is not None:
+                    data_copy['llm_config'] = llm_configs.get('llm')
+                config = create_condenser_config(condenser_type, data_copy)
+        else:
+            config = create_condenser_config(condenser_type, data)
+
+        condenser_mapping['condenser'] = config
+    except (ValidationError, ValueError) as e:
+        logger.openhands_logger.warning(
+            f'Invalid condenser configuration: {e}. Using NoOpCondenserConfig.'
+        )
+        # Default to NoOpCondenserConfig if config fails
+        config = NoOpCondenserConfig()
+        condenser_mapping['condenser'] = config
+
+    return condenser_mapping
+
+
+# For backward compatibility
+from_toml_section = condenser_config_from_toml_section
+
+
+def create_condenser_config(condenser_type: str, data: dict) -> CondenserConfig:
+    """
+    Create a CondenserConfig instance based on the specified type.
+
+    Args:
+        condenser_type: The type of condenser to create.
+        data: The configuration data.
+
+    Returns:
+        A CondenserConfig instance.
+
+    Raises:
+        ValueError: If the condenser type is unknown.
+        ValidationError: If the provided data fails validation for the condenser type.
+    """
+    # Mapping of condenser types to their config classes
+    condenser_classes = {
+        'noop': NoOpCondenserConfig,
+        'observation_masking': ObservationMaskingCondenserConfig,
+        'recent': RecentEventsCondenserConfig,
+        'llm': LLMSummarizingCondenserConfig,
+        'amortized': AmortizedForgettingCondenserConfig,
+        'llm_attention': LLMAttentionCondenserConfig,
+    }
+
+    if condenser_type not in condenser_classes:
+        raise ValueError(f'Unknown condenser type: {condenser_type}')
+
+    # Create and validate the config using direct instantiation
+    # Explicitly handle ValidationError to provide more context
+    try:
+        config_class = condenser_classes[condenser_type]
+        # Use type casting to help mypy understand the return type
+        return cast(CondenserConfig, config_class(**data))
+    except ValidationError as e:
+        # Just re-raise with a more descriptive message, but don't try to pass the errors
+        # which can cause compatibility issues with different pydantic versions
+        raise ValueError(
+            f"Validation failed for condenser type '{condenser_type}': {e}"
+        )

openhands/core/config/llm_config.py

@@ -48,7 +48,7 @@ class LLMConfig(BaseModel):
         reasoning_effort: The effort to put into reasoning. This is a string that can be one of 'low', 'medium', 'high', or 'none'. Exclusive for o1 models.
     """
 
-    model: str = Field(default='claude-3-5-sonnet-20241022')
+    model: str = Field(default='claude-3-7-sonnet-20250219')
     api_key: SecretStr | None = Field(default=None)
     base_url: str | None = Field(default=None)
     api_version: str | None = Field(default=None)

openhands/core/config/utils.py

@@ -12,9 +12,11 @@ import toml
 from dotenv import load_dotenv
 from pydantic import BaseModel, SecretStr, ValidationError
 
+from openhands import __version__
 from openhands.core import logger
 from openhands.core.config.agent_config import AgentConfig
 from openhands.core.config.app_config import AppConfig
+from openhands.core.config.condenser_config import condenser_config_from_toml_section
 from openhands.core.config.config_utils import (
     OH_DEFAULT_AGENT,
     OH_MAX_ITERATIONS,
@@ -193,6 +195,44 @@ def load_from_toml(cfg: AppConfig, toml_file: str = 'config.toml') -> None:
             # Re-raise ValueError from SandboxConfig.from_toml_section
             raise ValueError('Error in [sandbox] section in config.toml')
 
+    # Process condenser section if present
+    if 'condenser' in toml_config:
+        try:
+            # Pass the LLM configs to the condenser config parser
+            condenser_mapping = condenser_config_from_toml_section(
+                toml_config['condenser'], cfg.llms
+            )
+            # Assign the default condenser configuration to the default agent configuration
+            if 'condenser' in condenser_mapping:
+                # Get the default agent config and assign the condenser config to it
+                default_agent_config = cfg.get_agent_config()
+                default_agent_config.condenser = condenser_mapping['condenser']
+                logger.openhands_logger.debug(
+                    'Default condenser configuration loaded from config toml and assigned to default agent'
+                )
+        except (TypeError, KeyError, ValidationError) as e:
+            logger.openhands_logger.warning(
+                f'Cannot parse [condenser] config from toml, values have not been applied.\nError: {e}'
+            )
+    # If no condenser section is in toml but enable_default_condenser is True,
+    # set LLMSummarizingCondenserConfig as default
+    elif cfg.enable_default_condenser:
+        from openhands.core.config.condenser_config import LLMSummarizingCondenserConfig
+
+        # Get default agent config
+        default_agent_config = cfg.get_agent_config()
+
+        # Create default LLM summarizing condenser config
+        default_condenser = LLMSummarizingCondenserConfig(
+            llm_config=cfg.get_llm_config(),  # Use default LLM config
+        )
+
+        # Set as default condenser
+        default_agent_config.condenser = default_condenser
+        logger.openhands_logger.debug(
+            'Default LLM summarizing condenser assigned to default agent (no condenser in config)'
+        )
+
     # Process extended section if present
     if 'extended' in toml_config:
         try:
@@ -203,7 +243,15 @@ def load_from_toml(cfg: AppConfig, toml_file: str = 'config.toml') -> None:
             )
 
     # Check for unknown sections
-    known_sections = {'core', 'extended', 'agent', 'llm', 'security', 'sandbox'}
+    known_sections = {
+        'core',
+        'extended',
+        'agent',
+        'llm',
+        'security',
+        'sandbox',
+        'condenser',
+    }
     for key in toml_config:
         if key.lower() not in known_sections:
             logger.openhands_logger.warning(f'Unknown section [{key}] in {toml_file}')
@@ -492,8 +540,6 @@ def parse_arguments() -> argparse.Namespace:
     args = parser.parse_args()
 
     if args.version:
-        from openhands import __version__
-
         print(f'OpenHands version: {__version__}')
         sys.exit(0)
 

openhands/core/main.py

@@ -20,6 +20,7 @@ from openhands.core.setup import (
     create_controller,
     create_runtime,
     generate_sid,
+    initialize_repository_for_runtime,
 )
 from openhands.events import EventSource, EventStreamSubscriber
 from openhands.events.action import MessageAction, NullAction
@@ -29,6 +30,7 @@ from openhands.events.observation import AgentStateChangedObservation
 from openhands.events.serialization import event_from_dict
 from openhands.io import read_input, read_task
 from openhands.runtime.base import Runtime
+from openhands.utils.async_utils import call_async_from_sync
 
 
 class FakeUserResponseFunc(Protocol):
@@ -97,8 +99,17 @@ async def run_controller(
             sid=sid,
             headless_mode=headless_mode,
             agent=agent,
-            selected_repository=config.sandbox.selected_repo,
         )
+        # Connect to the runtime
+        call_async_from_sync(runtime.connect)
+
+        # Initialize repository if needed
+        if config.sandbox.selected_repo:
+            initialize_repository_for_runtime(
+                runtime,
+                agent=agent,
+                selected_repository=config.sandbox.selected_repo,
+            )
 
     event_stream = runtime.event_stream
 

openhands/core/message_format.md

@@ -7,6 +7,7 @@ OpenHands uses its own `Message` class (`openhands/core/message.py`) which provi
 ## Class Structure
 
 Our `Message` class (`openhands/core/message.py`):
+
 ```python
 class Message(BaseModel):
     role: Literal['user', 'system', 'assistant', 'tool']
@@ -22,13 +23,14 @@ class Message(BaseModel):
 ```
 
 litellm's `Message` class (`litellm/types/utils.py`):
+
 ```python
 class Message(OpenAIObject):
-    content: Optional[str]
+    content: str | None
     role: Literal["assistant", "user", "system", "tool", "function"]
-    tool_calls: Optional[List[ChatCompletionMessageToolCall]]
-    function_call: Optional[FunctionCall]
-    audio: Optional[ChatCompletionAudioResponse] = None
+    tool_calls: List[ChatCompletionMessageToolCall] | None
+    function_call: FunctionCall | None
+    audio: ChatCompletionAudioResponse | None = None
 ```
 
 ## How It Works
@@ -36,6 +38,7 @@ class Message(OpenAIObject):
 1. **Message Creation**: Our `Message` class is a Pydantic model that supports rich content (text and images) through its `content` field.
 
 2. **Serialization**: The class uses Pydantic's `@model_serializer` to convert messages into dictionaries that litellm can understand. We have two serialization methods:
+
    ```python
    def _string_serializer(self) -> dict:
        # convert content to a single string
@@ -55,6 +58,7 @@ class Message(OpenAIObject):
    ```
 
    The appropriate serializer is chosen based on the message's capabilities:
+
    ```python
    @model_serializer
    def serialize_model(self) -> dict:
@@ -64,11 +68,13 @@ class Message(OpenAIObject):
    ```
 
 3. **Tool Call Handling**: Tool calls require special attention in serialization because:
+
    - They need to work with litellm's API calls (which accept both dicts and objects)
    - They need to be properly serialized for token counting
    - They need to maintain compatibility with different LLM providers' formats
 
 4. **litellm Integration**: When we pass our messages to `litellm.completion()`, litellm doesn't care about the message class type - it works with the dictionary representation. This works because:
+
    - litellm's transformation code (e.g., `litellm/llms/anthropic/chat/transformation.py`) processes messages based on their structure, not their type
    - our serialization produces dictionaries that match litellm's expected format
    - litellm handles rich content by looking at the message structure, supporting both simple string content and lists of content items
@@ -78,6 +84,7 @@ class Message(OpenAIObject):
 ### Token Counting
 
 To use litellm's token counter, we need to make sure that all message components (including tool calls) are properly serialized to dictionaries. This is because:
+
 - litellm's token counter expects dictionary structures
 - Tool calls need to be included in the token count
 - Different providers may count tokens differently for structured content

openhands/core/setup.py

@@ -21,7 +21,6 @@ from openhands.runtime import get_runtime_cls
 from openhands.runtime.base import Runtime
 from openhands.security import SecurityAnalyzer, options
 from openhands.storage import get_file_store
-from openhands.utils.async_utils import call_async_from_sync
 
 
 def create_runtime(
@@ -29,18 +28,19 @@ def create_runtime(
     sid: str | None = None,
     headless_mode: bool = True,
     agent: Agent | None = None,
-    selected_repository: str | None = None,
-    github_token: SecretStr | None = None,
 ) -> Runtime:
     """Create a runtime for the agent to run on.
 
-    config: The app config.
-    sid: (optional) The session id. IMPORTANT: please don't set this unless you know what you're doing.
-        Set it to incompatible value will cause unexpected behavior on RemoteRuntime.
-    headless_mode: Whether the agent is run in headless mode. `create_runtime` is typically called within evaluation scripts,
-        where we don't want to have the VSCode UI open, so it defaults to True.
-    selected_repository: (optional) The GitHub repository to use.
-    github_token: (optional) The GitHub token to use.
+    Args:
+        config: The app config.
+        sid: (optional) The session id. IMPORTANT: please don't set this unless you know what you're doing.
+            Set it to incompatible value will cause unexpected behavior on RemoteRuntime.
+        headless_mode: Whether the agent is run in headless mode. `create_runtime` is typically called within evaluation scripts,
+            where we don't want to have the VSCode UI open, so it defaults to True.
+        agent: (optional) The agent instance to use for configuring the runtime.
+
+    Returns:
+        The created Runtime instance (not yet connected or initialized).
     """
     # if sid is provided on the command line, use it as the name of the event stream
     # otherwise generate it on the basis of the configured jwt_secret
@@ -74,8 +74,30 @@ def create_runtime(
         headless_mode=headless_mode,
     )
 
-    call_async_from_sync(runtime.connect)
+    logger.debug(
+        f'Runtime created with plugins: {[plugin.name for plugin in runtime.plugins]}'
+    )
 
+    return runtime
+
+
+def initialize_repository_for_runtime(
+    runtime: Runtime,
+    agent: Agent | None = None,
+    selected_repository: str | None = None,
+    github_token: SecretStr | None = None,
+) -> str | None:
+    """Initialize the repository for the runtime.
+
+    Args:
+        runtime: The runtime to initialize the repository for.
+        agent: (optional) The agent to load microagents for.
+        selected_repository: (optional) The GitHub repository to use.
+        github_token: (optional) The GitHub token to use.
+
+    Returns:
+        The repository directory path if a repository was cloned, None otherwise.
+    """
     # clone selected repository if provided
     repo_directory = None
     github_token = (
@@ -98,11 +120,7 @@ def create_runtime(
         agent.prompt_manager.load_microagents(microagents)
         agent.prompt_manager.set_repository_info(selected_repository, repo_directory)
 
-    logger.debug(
-        f'Runtime initialized with plugins: {[plugin.name for plugin in runtime.plugins]}'
-    )
-
-    return runtime
+    return repo_directory
 
 
 def create_agent(config: AppConfig) -> Agent:

openhands/events/stream.py

@@ -428,3 +428,41 @@ class EventStream:
                 break
 
         return matching_events
+
+    def get_metrics(self):
+        """Get the accumulated metrics from all events in the stream.
+
+        This method extracts metrics from events that contain them and returns
+        the aggregated metrics object.
+
+        Returns:
+            Metrics: The metrics object containing accumulated cost and token usage data.
+            Returns None if no metrics are found.
+        """
+        from openhands.llm.metrics import Metrics
+
+        # Look for events with metrics
+        metrics = None
+        events_with_metrics = []
+
+        try:
+            # First collect all events with metrics
+            for event in self.get_events():
+                if hasattr(event, 'llm_metrics') and event.llm_metrics is not None:
+                    events_with_metrics.append(event)
+
+            # Then merge them if any were found
+            if events_with_metrics:
+                # Get the first event with metrics to initialize our metrics object
+                first_event = events_with_metrics[0]
+                if first_event.llm_metrics is not None:
+                    metrics = Metrics(model_name=first_event.llm_metrics.model_name)
+
+                    # Merge metrics from all events
+                    for event in events_with_metrics:
+                        if event.llm_metrics is not None:
+                            metrics.merge(event.llm_metrics)
+        except Exception as e:
+            logger.error(f'Error retrieving metrics from events: {e}')
+
+        return metrics

openhands/integrations/github/github_service.py

@@ -51,6 +51,9 @@ class GitHubService:
     async def get_latest_token(self) -> SecretStr:
         return self.token
 
+    async def get_latest_provider_token(self) -> SecretStr:
+        return self.token
+
     async def _fetch_data(
         self, url: str, params: dict | None = None
     ) -> tuple[Any, dict]:

openhands/llm/llm.py

@@ -76,6 +76,8 @@ REASONING_EFFORT_SUPPORTED_MODELS = [
 MODELS_WITHOUT_STOP_WORDS = [
     'o1-mini',
     'o1-preview',
+    'o1',
+    'o1-2024-12-17',
 ]
 
 
@@ -217,9 +219,8 @@ class LLM(RetryMixin, DebugMixin):
                     kwargs['stop'] = STOP_WORDS
 
                 mock_fncall_tools = kwargs.pop('tools')
-                kwargs['tool_choice'] = (
-                    'none'  # force no tool calling because we're mocking it - without it, it will cause issue with sglang
-                )
+                # tool_choice should not be specified when mocking function calling
+                kwargs.pop('tool_choice', None)
 
             # if we have no messages, something went very wrong
             if not messages:

openhands/memory/condenser/__init__.py

@@ -1,4 +1,4 @@
 import openhands.memory.condenser.impl  # noqa F401 (we import this to get the condensers registered)
 from openhands.memory.condenser.condenser import Condenser, get_condensation_metadata
 
-__all__ = ['Condenser', 'get_condensation_metadata']
+__all__ = ['Condenser', 'get_condensation_metadata', 'CONDENSER_REGISTRY']

openhands/resolver/resolve_issue.py

@@ -202,7 +202,7 @@ async def process_issue(
         timeout=300,
     )
 
-    if os.getenv('GITLAB_CI') == 'True':
+    if os.getenv('GITLAB_CI') == 'true':
         sandbox_config.local_runtime_url = os.getenv(
             'LOCAL_RUNTIME_URL', 'http://localhost'
         )
@@ -651,7 +651,7 @@ def main() -> None:
     if not token:
         raise ValueError('Token is required.')
 
-    platform = identify_token(token)
+    platform = identify_token(token, repo)
     if platform == Platform.INVALID:
         raise ValueError('Token is invalid.')
 

openhands/resolver/utils.py

@@ -22,18 +22,37 @@ class Platform(Enum):
     GITLAB = 2
 
 
-def identify_token(token: str) -> Platform:
+def identify_token(token: str, repo: str | None = None) -> Platform:
     """
     Identifies whether a token belongs to GitHub or GitLab.
 
     Parameters:
         token (str): The personal access token to check.
+        repo (str): Repository in format "owner/repo" for GitHub Actions token validation.
 
     Returns:
         Platform: "GitHub" if the token is valid for GitHub,
              "GitLab" if the token is valid for GitLab,
              "Invalid" if the token is not recognized by either.
     """
+    # Try GitHub Actions token format (Bearer) with repo endpoint if repo is provided
+    if repo:
+        github_repo_url = f'https://api.github.com/repos/{repo}'
+        github_bearer_headers = {
+            'Authorization': f'Bearer {token}',
+            'Accept': 'application/vnd.github+json',
+        }
+
+        try:
+            github_repo_response = requests.get(
+                github_repo_url, headers=github_bearer_headers, timeout=5
+            )
+            if github_repo_response.status_code == 200:
+                return Platform.GITHUB
+        except requests.RequestException as e:
+            print(f'Error connecting to GitHub API (repo check): {e}')
+
+    # Try GitHub PAT format (token)
     github_url = 'https://api.github.com/user'
     github_headers = {'Authorization': f'token {token}'}
 
@@ -44,6 +63,7 @@ def identify_token(token: str) -> Platform:
     except requests.RequestException as e:
         print(f'Error connecting to GitHub API: {e}')
 
+    # Try GitLab token
     gitlab_url = 'https://gitlab.com/api/v4/user'
     gitlab_headers = {'Authorization': f'Bearer {token}'}
 

openhands/runtime/base.py

@@ -222,7 +222,7 @@ class Runtime(FileEditRuntimeMixin):
             if isinstance(event, CmdRunAction):
                 if self.github_user_id and '$GITHUB_TOKEN' in event.command:
                     gh_client = GithubServiceImpl(user_id=self.github_user_id)
-                    token = await gh_client.get_latest_token()
+                    token = await gh_client.get_latest_provider_token()
                     if token:
                         export_cmd = CmdRunAction(
                             f"export GITHUB_TOKEN='{token.get_secret_value()}'"

openhands/runtime/builder/docker.py

@@ -19,16 +19,19 @@ class DockerRuntimeBuilder(RuntimeBuilder):
 
         version_info = self.docker_client.version()
         server_version = version_info.get('Version', '').replace('-', '.')
-        self.is_podman = version_info.get('Components')[0].get('Name').startswith('Podman')
-        if tuple(map(int, server_version.split('.')[:2])) < (18, 9) and not self.is_podman:
+        self.is_podman = (
+            version_info.get('Components')[0].get('Name').startswith('Podman')
+        )
+        if (
+            tuple(map(int, server_version.split('.')[:2])) < (18, 9)
+            and not self.is_podman
+        ):
             raise AgentRuntimeBuildError(
                 'Docker server version must be >= 18.09 to use BuildKit'
             )
 
         if self.is_podman and tuple(map(int, server_version.split('.')[:2])) < (4, 9):
-            raise AgentRuntimeBuildError(
-                'Podman server version must be >= 4.9.0'
-            )
+            raise AgentRuntimeBuildError('Podman server version must be >= 4.9.0')
 
         self.rolling_logger = RollingLogger(max_lines=10)
 
@@ -37,7 +40,9 @@ class DockerRuntimeBuilder(RuntimeBuilder):
         """Check if Docker Buildx is available"""
         try:
             result = subprocess.run(
-                ['docker' if not is_podman else 'podman', 'buildx', 'version'], capture_output=True, text=True
+                ['docker' if not is_podman else 'podman', 'buildx', 'version'],
+                capture_output=True,
+                text=True,
             )
             return result.returncode == 0
         except FileNotFoundError:
@@ -74,16 +79,16 @@ class DockerRuntimeBuilder(RuntimeBuilder):
         self.docker_client = docker.from_env()
         version_info = self.docker_client.version()
         server_version = version_info.get('Version', '').split('+')[0].replace('-', '.')
-        self.is_podman = version_info.get('Components')[0].get('Name').startswith('Podman')
+        self.is_podman = (
+            version_info.get('Components')[0].get('Name').startswith('Podman')
+        )
         if tuple(map(int, server_version.split('.'))) < (18, 9) and not self.is_podman:
             raise AgentRuntimeBuildError(
                 'Docker server version must be >= 18.09 to use BuildKit'
             )
 
         if self.is_podman and tuple(map(int, server_version.split('.'))) < (4, 9):
-            raise AgentRuntimeBuildError(
-                'Podman server version must be >= 4.9.0'
-            )
+            raise AgentRuntimeBuildError('Podman server version must be >= 4.9.0')
 
         if not DockerRuntimeBuilder.check_buildx(self.is_podman):
             # when running openhands in a container, there might not be a "docker"

openhands/runtime/impl/daytona/README.md

@@ -2,22 +2,80 @@
 
 [Daytona](https://www.daytona.io/) is a platform that provides a secure and elastic infrastructure for running AI-generated code. It provides all the necessary features for an AI Agent to interact with a codebase. It provides a Daytona SDK with official Python and TypeScript interfaces for interacting with Daytona, enabling you to programmatically manage development environments and execute code.
 
-## Getting started
+## Quick Start
 
-1. Sign in at https://app.daytona.io/
-
-1. Generate and copy your API key
-
-1. Set the following environment variables before running the OpenHands app on your local machine or via a `docker run` command:
+### Step 1: Retrieve Your Daytona API Key
+1. Visit the [Daytona Dashboard](https://app.daytona.io/dashboard/keys).
+2. Click **"Create Key"**.
+3. Enter a name for your key and confirm the creation.
+4. Once the key is generated, copy it.
 
+### Step 2: Set Your API Key as an Environment Variable
+Run the following command in your terminal, replacing `<your-api-key>` with the actual key you copied:
 ```bash
-    RUNTIME="daytona"
-    DAYTONA_API_KEY="<your-api-key>"
+export DAYTONA_API_KEY="<your-api-key>"
 ```
-Optionally, if you don't want your sandboxes to default to the US region, set:
+
+This step ensures that OpenHands can authenticate with the Daytona platform when it runs.
+
+### Step 3: Run OpenHands Locally Using Docker
+To start the latest version of OpenHands on your machine, execute the following command in your terminal:
+```bash
+bash -i <(curl -sL https://get.daytona.io/openhands)
+```
+
+#### What This Command Does:
+- Downloads the latest OpenHands release script.
+- Runs the script in an interactive Bash session.
+- Automatically pulls and runs the OpenHands container using Docker.
+Once executed, OpenHands should be running locally and ready for use.
+
+
+## Manual Initialization
+
+### Step 1: Set the `OPENHANDS_VERSION` Environment Variable
+Run the following command in your terminal, replacing `<openhands-release>` with the latest release's version seen in the [main README.md file](https://github.com/All-Hands-AI/OpenHands?tab=readme-ov-file#-quick-start):
 
 ```bash
-    DAYTONA_TARGET="eu"
+export OPENHANDS_VERSION="<openhands-release>"  # e.g. 0.27
+```
+
+### Step 2: Retrieve Your Daytona API Key
+1. Visit the [Daytona Dashboard](https://app.daytona.io/dashboard/keys).
+2. Click **"Create Key"**.
+3. Enter a name for your key and confirm the creation.
+4. Once the key is generated, copy it.
+
+### Step 3: Set Your API Key as an Environment Variable:
+Run the following command in your terminal, replacing `<your-api-key>` with the actual key you copied:
+```bash
+export DAYTONA_API_KEY="<your-api-key>"
+```
+
+### Step 4: Run the following `docker` command:
+This command pulls and runs the OpenHands container using Docker. Once executed, OpenHands should be running locally and ready for use.
+
+```bash
+docker run -it --rm --pull=always \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:${OPENHANDS_VERSION}-nikolaik \
+    -e LOG_ALL_EVENTS=true \
+    -e RUNTIME=daytona \
+    -e DAYTONA_API_KEY=${DAYTONA_API_KEY} \
+    -v ~/.openhands-state:/.openhands-state \
+    -p 3000:3000 \
+    --name openhands-app \
+    docker.all-hands.dev/all-hands-ai/openhands:${OPENHANDS_VERSION}
+```
+
+> **Tip:** If you don't want your sandboxes to default to the US region, you can set the `DAYTONA_TARGET` environment variable to `eu`
+
+### Running OpenHands Locally Without Docker
+
+Alternatively, if you want to run the OpenHands app on your local machine using `make run` without Docker, make sure to set the following environment variables first:
+
+```bash
+export RUNTIME="daytona"
+export DAYTONA_API_KEY="<your-api-key>"
 ```
 
 ## Documentation

openhands/runtime/impl/docker/docker_runtime.py

@@ -333,7 +333,10 @@ class DockerRuntime(ActionExecutionClient):
         if exposed_ports:
             for exposed_port in exposed_ports.keys():
                 exposed_port = int(exposed_port.split('/tcp')[0])
-                if exposed_port != self._host_port and exposed_port != self._vscode_port:
+                if (
+                    exposed_port != self._host_port
+                    and exposed_port != self._vscode_port
+                ):
                     self._app_ports.append(exposed_port)
 
         self.api_url = f'{self.config.sandbox.local_runtime_url}:{self._container_port}'

openhands/runtime/impl/e2b/e2b_runtime.py

@@ -1,4 +1,4 @@
-from typing import Callable, Optional
+from typing import Callable
 
 from openhands.core.config import AppConfig
 from openhands.events.action import (
@@ -27,7 +27,7 @@ class E2BRuntime(Runtime):
         sid: str = 'default',
         plugins: list[PluginRequirement] | None = None,
         sandbox: E2BSandbox | None = None,
-        status_callback: Optional[Callable] = None,
+        status_callback: Callable | None = None,
     ):
         super().__init__(
             config,

openhands/runtime/impl/local/local_runtime.py

@@ -7,7 +7,7 @@ import shutil
 import subprocess
 import tempfile
 import threading
-from typing import Callable, Optional
+from typing import Callable
 
 import requests
 import tenacity
@@ -155,7 +155,7 @@ class LocalRuntime(ActionExecutionClient):
 
         self.api_url = f'{self.config.sandbox.local_runtime_url}:{self._host_port}'
         self.status_callback = status_callback
-        self.server_process: Optional[subprocess.Popen[str]] = None
+        self.server_process: subprocess.Popen[str] | None = None
         self.action_semaphore = threading.Semaphore(1)  # Ensure one action at a time
 
         # Update env vars

openhands/runtime/impl/remote/remote_runtime.py

@@ -1,5 +1,5 @@
 import os
-from typing import Callable, Optional
+from typing import Callable
 from urllib.parse import urlparse
 
 import requests
@@ -42,7 +42,7 @@ class RemoteRuntime(ActionExecutionClient):
         sid: str = 'default',
         plugins: list[PluginRequirement] | None = None,
         env_vars: dict[str, str] | None = None,
-        status_callback: Optional[Callable] = None,
+        status_callback: Callable | None = None,
         attach_to_existing: bool = False,
         headless_mode: bool = True,
         github_user_id: str | None = None,

openhands/runtime/utils/vscode-extensions/memory-monitor/README.md

@@ -45,4 +45,4 @@ This extension is part of the OpenHands project. To modify or extend it:
 
 ## License
 
-This extension is licensed under the MIT license.
+This extension is licensed under the MIT license.

openhands/runtime/utils/vscode-extensions/memory-monitor/extension.js

@@ -4,30 +4,30 @@ const MemoryMonitor = require('./memory_monitor');
 function activate(context) {
     // Create memory monitor instance
     const memoryMonitor = new MemoryMonitor();
-    
+
     // Store the context in the memory monitor
     memoryMonitor.context = context;
-    
+
     // Register memory monitor start command
     let startMonitorCommand = vscode.commands.registerCommand('openhands-memory-monitor.startMemoryMonitor', function () {
         memoryMonitor.start();
     });
-    
+
     // Register memory monitor stop command
     let stopMonitorCommand = vscode.commands.registerCommand('openhands-memory-monitor.stopMemoryMonitor', function () {
         memoryMonitor.stop();
     });
-    
+
     // Register memory details command
     let showMemoryDetailsCommand = vscode.commands.registerCommand('openhands-memory-monitor.showMemoryDetails', function () {
         memoryMonitor.showDetails();
     });
-    
+
     // Add all commands to subscriptions
     context.subscriptions.push(startMonitorCommand);
     context.subscriptions.push(stopMonitorCommand);
     context.subscriptions.push(showMemoryDetailsCommand);
-    
+
     // Start memory monitoring by default
     memoryMonitor.start();
 }
@@ -39,4 +39,4 @@ function deactivate() {
 module.exports = {
     activate,
     deactivate
-}
+}

openhands/runtime/utils/vscode-extensions/memory-monitor/memory_monitor.js

@@ -21,15 +21,15 @@ class MemoryMonitor {
 
         this.isMonitoring = true;
         this.statusBarItem.show();
-        
+
         // Initial update
         this.updateMemoryInfo();
-        
+
         // Set interval for updates
         this.intervalId = setInterval(() => {
             this.updateMemoryInfo();
         }, interval);
-        
+
         vscode.window.showInformationMessage('Memory monitoring started');
     }
 
@@ -41,7 +41,7 @@ class MemoryMonitor {
         this.isMonitoring = false;
         clearInterval(this.intervalId);
         this.statusBarItem.hide();
-        
+
         vscode.window.showInformationMessage('Memory monitoring stopped');
     }
 
@@ -49,18 +49,18 @@ class MemoryMonitor {
         const totalMem = os.totalmem();
         const freeMem = os.freemem();
         const usedMem = totalMem - freeMem;
-        
+
         // Calculate memory usage percentage
         const memUsagePercent = Math.round((usedMem / totalMem) * 100);
-        
+
         // Format memory values to MB
         const usedMemMB = Math.round(usedMem / (1024 * 1024));
         const totalMemMB = Math.round(totalMem / (1024 * 1024));
-        
+
         // Update status bar
         this.statusBarItem.text = `$(pulse) Mem: ${memUsagePercent}%`;
         this.statusBarItem.tooltip = `Memory Usage: ${usedMemMB}MB / ${totalMemMB}MB`;
-        
+
         // Store memory data in history
         this.memoryHistory.push({
             timestamp: new Date(),
@@ -69,7 +69,7 @@ class MemoryMonitor {
             memUsagePercent,
             processMemory: process.memoryUsage()
         });
-        
+
         // Limit history length
         if (this.memoryHistory.length > this.maxHistoryLength) {
             this.memoryHistory.shift();
@@ -86,7 +86,7 @@ class MemoryMonitor {
                 enableScripts: true
             }
         );
-        
+
         // Set up message handler for real-time updates
         panel.webview.onDidReceiveMessage(
             message => {
@@ -97,60 +97,60 @@ class MemoryMonitor {
             undefined,
             this.context ? this.context.subscriptions : []
         );
-        
+
         // Initial update
         this.updateWebviewContent(panel);
-        
+
         // Handle panel disposal
         panel.onDidDispose(() => {
             // Clean up any resources if needed
         }, null, this.context ? this.context.subscriptions : []);
     }
-    
+
     updateWebviewContent(panel) {
         // Get system memory info
         const totalMem = os.totalmem();
         const freeMem = os.freemem();
         const usedMem = totalMem - freeMem;
-        
+
         // Format memory values
         const usedMemMB = Math.round(usedMem / (1024 * 1024));
         const freeMemMB = Math.round(freeMem / (1024 * 1024));
         const totalMemMB = Math.round(totalMem / (1024 * 1024));
-        
+
         // Get process memory usage
         const processMemory = process.memoryUsage();
         const rss = Math.round(processMemory.rss / (1024 * 1024));
         const heapTotal = Math.round(processMemory.heapTotal / (1024 * 1024));
         const heapUsed = Math.round(processMemory.heapUsed / (1024 * 1024));
-        
+
         // Get process information
         this.processMonitor.getProcessInfo((error, processInfo) => {
             if (error) {
                 console.error('Error getting process info:', error);
                 return;
             }
-            
+
             // Create HTML content for the webview
             const htmlContent = this.generateHtmlReport(
-                usedMemMB, freeMemMB, totalMemMB, 
+                usedMemMB, freeMemMB, totalMemMB,
                 rss, heapTotal, heapUsed,
                 processInfo
             );
-            
+
             // Set the webview's HTML content
             panel.webview.html = htmlContent;
         });
     }
-    
+
     generateHtmlReport(usedMemMB, freeMemMB, totalMemMB, rss, heapTotal, heapUsed, processInfo) {
         // Create memory usage history data for chart
         const memoryLabels = this.memoryHistory.map((entry, index) => index);
         const memoryData = this.memoryHistory.map(entry => entry.memUsagePercent);
-        const heapData = this.memoryHistory.map(entry => 
+        const heapData = this.memoryHistory.map(entry =>
             Math.round(entry.processMemory.heapUsed / (1024 * 1024))
         );
-        
+
         // Format process info table
         let processTable = '';
         if (processInfo && processInfo.processes) {
@@ -174,7 +174,7 @@ class MemoryMonitor {
                 </table>
             `;
         }
-        
+
         return `
             <!DOCTYPE html>
             <html lang="en">
@@ -237,7 +237,7 @@ class MemoryMonitor {
             </head>
             <body>
                 <h1>Memory Monitor</h1>
-                
+
                 <div class="memory-card">
                     <h2>System Memory</h2>
                     <div class="memory-info">
@@ -259,7 +259,7 @@ class MemoryMonitor {
                         </div>
                     </div>
                 </div>
-                
+
                 <div class="memory-card">
                     <h2>Process Memory (VSCode Extension Host)</h2>
                     <div class="memory-info">
@@ -277,18 +277,18 @@ class MemoryMonitor {
                         </div>
                     </div>
                 </div>
-                
+
                 <div class="memory-card">
                     <h2>Memory Usage History</h2>
                     <div class="chart-container">
                         <canvas id="memoryChart"></canvas>
                     </div>
                 </div>
-                
+
                 <div class="memory-card">
                     ${processTable}
                 </div>
-                
+
                 <script>
                     // Create memory usage chart
                     const ctx = document.getElementById('memoryChart').getContext('2d');
@@ -323,10 +323,10 @@ class MemoryMonitor {
                             }
                         }
                     });
-                    
+
                     // Set up real-time updates
                     const vscode = acquireVsCodeApi();
-                    
+
                     // Request updates every 5 seconds
                     setInterval(() => {
                         vscode.postMessage({
@@ -340,4 +340,4 @@ class MemoryMonitor {
     }
 }
 
-module.exports = MemoryMonitor;
+module.exports = MemoryMonitor;

openhands/runtime/utils/vscode-extensions/memory-monitor/package.json

@@ -47,4 +47,4 @@
             ]
         }
     }
-}
+}

openhands/runtime/utils/vscode-extensions/memory-monitor/process_monitor.js

@@ -44,7 +44,7 @@ class ProcessMonitor {
                 const memPercent = parseFloat(parts[parts.length - 2]);
                 const cpuPercent = parseFloat(parts[parts.length - 1]);
                 const cmd = parts.slice(2, parts.length - 2).join(' ');
-                
+
                 return {
                     pid,
                     ppid,
@@ -80,7 +80,7 @@ class ProcessMonitor {
                 const memPercent = parseFloat(parts[parts.length - 2]);
                 const cpuPercent = parseFloat(parts[parts.length - 1]);
                 const cmd = parts.slice(2, parts.length - 2).join(' ');
-                
+
                 return {
                     pid,
                     ppid,
@@ -109,21 +109,21 @@ class ProcessMonitor {
             // Parse the CSV output
             const lines = stdout.trim().split('\n');
             const header = "PID,PPID,Command,Memory (bytes)";
-            
+
             // Skip empty lines and the header
             const dataLines = lines.filter(line => line.trim() !== '' && !line.includes('Node,'));
-            
+
             const processes = dataLines.map(line => {
                 const parts = line.split(',');
                 if (parts.length < 4) return null;
-                
+
                 // Last part is the node name, then ProcessId, ParentProcessId, CommandLine, WorkingSetSize
                 const pid = parts[parts.length - 4];
                 const ppid = parts[parts.length - 3];
                 const cmd = parts[parts.length - 2];
                 const memBytes = parseInt(parts[parts.length - 1], 10);
                 const memPercent = (memBytes / os.totalmem() * 100).toFixed(1);
-                
+
                 return {
                     pid,
                     ppid,
@@ -140,4 +140,4 @@ class ProcessMonitor {
     }
 }
 
-module.exports = ProcessMonitor;
+module.exports = ProcessMonitor;

openhands/server/data_models/feedback.py

@@ -1,5 +1,5 @@
 import json
-from typing import Any, Literal, Optional
+from typing import Any, Literal
 
 import requests
 from pydantic import BaseModel
@@ -15,7 +15,7 @@ class FeedbackDataModel(BaseModel):
         'positive', 'negative'
     ]  # TODO: remove this, its here for backward compatibility
     permissions: Literal['public', 'private']
-    trajectory: Optional[list[dict[str, Any]]]
+    trajectory: list[dict[str, Any]] | None
 
 
 FEEDBACK_URL = 'https://share-od-trajectory-3u9bw9tx.uc.gateway.dev/share_od_trajectory'

openhands/server/routes/conversation.py

@@ -7,6 +7,70 @@ from openhands.runtime.base import Runtime
 app = APIRouter(prefix='/api/conversations/{conversation_id}')
 
 
+@app.get('/metrics')
+async def get_conversation_metrics(request: Request):
+    """Retrieve the conversation metrics.
+
+    This endpoint returns the accumulated cost and token usage metrics for the conversation.
+
+    Args:
+        request (Request): The incoming FastAPI request object.
+
+    Returns:
+        JSONResponse: A JSON response containing the metrics data.
+    """
+    try:
+        if not hasattr(request.state, 'conversation'):
+            return JSONResponse(
+                status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+                content={'error': 'No conversation found in request state'},
+            )
+
+        event_stream = request.state.conversation.event_stream
+
+        # Get metrics from the event stream
+        metrics = (
+            event_stream.get_metrics() if hasattr(event_stream, 'get_metrics') else None
+        )
+
+        if not metrics:
+            # Return empty metrics if not available
+            return JSONResponse(
+                status_code=status.HTTP_200_OK,
+                content={
+                    'accumulated_cost': 0.0,
+                    'total_prompt_tokens': 0,
+                    'total_completion_tokens': 0,
+                    'total_tokens': 0,
+                },
+            )
+
+        # Calculate total tokens
+        total_prompt_tokens = sum(usage.prompt_tokens for usage in metrics.token_usages)
+        total_completion_tokens = sum(
+            usage.completion_tokens for usage in metrics.token_usages
+        )
+        total_tokens = total_prompt_tokens + total_completion_tokens
+
+        return JSONResponse(
+            status_code=status.HTTP_200_OK,
+            content={
+                'accumulated_cost': metrics.accumulated_cost,
+                'total_prompt_tokens': total_prompt_tokens,
+                'total_completion_tokens': total_completion_tokens,
+                'total_tokens': total_tokens,
+            },
+        )
+    except Exception as e:
+        logger.error(f'Error getting conversation metrics: {e}')
+        return JSONResponse(
+            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
+            content={
+                'error': f'Error getting conversation metrics: {e}',
+            },
+        )
+
+
 @app.get('/config')
 async def get_remote_runtime_config(request: Request):
     """Retrieve the runtime configuration.

openhands/server/routes/feedback.py

@@ -11,7 +11,7 @@ app = APIRouter(prefix='/api/conversations/{conversation_id}')
 
 
 @app.post('/submit-feedback')
-async def submit_feedback(request: Request, conversation_id: str):
+async def submit_feedback(request: Request, conversation_id: str) -> JSONResponse:
     """Submit user feedback.
 
     This function stores the provided feedback data.

openhands/server/routes/github.py

@@ -15,7 +15,7 @@ from openhands.server.auth import get_github_token, get_idp_token, get_user_id
 app = APIRouter(prefix='/api/github')
 
 
-@app.get('/repositories')
+@app.get('/repositories', response_model=list[GitHubRepository])
 async def get_github_repositories(
     page: int = 1,
     per_page: int = 10,
@@ -47,7 +47,7 @@ async def get_github_repositories(
         )
 
 
-@app.get('/user')
+@app.get('/user', response_model=GitHubUser)
 async def get_github_user(
     github_user_id: str | None = Depends(get_user_id),
     github_user_token: SecretStr | None = Depends(get_github_token),
@@ -73,7 +73,7 @@ async def get_github_user(
         )
 
 
-@app.get('/installations')
+@app.get('/installations', response_model=list[int])
 async def get_github_installation_ids(
     github_user_id: str | None = Depends(get_user_id),
     github_user_token: SecretStr | None = Depends(get_github_token),
@@ -99,7 +99,7 @@ async def get_github_installation_ids(
         )
 
 
-@app.get('/search/repositories')
+@app.get('/search/repositories', response_model=list[GitHubRepository])
 async def search_github_repositories(
     query: str,
     per_page: int = 5,
@@ -131,7 +131,7 @@ async def search_github_repositories(
         )
 
 
-@app.get('/suggested-tasks')
+@app.get('/suggested-tasks', response_model=list[SuggestedTask])
 async def get_suggested_tasks(
     github_user_id: str | None = Depends(get_user_id),
     github_user_token: SecretStr | None = Depends(get_github_token),

openhands/server/routes/manage_conversations.py

@@ -51,6 +51,7 @@ async def _create_new_conversation(
     selected_branch: str | None,
     initial_user_msg: str | None,
     image_urls: list[str] | None,
+    attach_convo_id: bool = False,
 ):
     monitoring_listener.on_create_conversation()
     logger.info('Loading settings')
@@ -109,8 +110,13 @@ async def _create_new_conversation(
     logger.info(f'Starting agent loop for conversation {conversation_id}')
     initial_message_action = None
     if initial_user_msg or image_urls:
+        user_msg = (
+            initial_user_msg.format(conversation_id)
+            if attach_convo_id and initial_user_msg
+            else initial_user_msg
+        )
         initial_message_action = MessageAction(
-            content=initial_user_msg or '',
+            content=user_msg or '',
             image_urls=image_urls or [],
         )
     event_stream = await conversation_manager.maybe_start_agent_loop(

openhands/server/routes/public.py

@@ -1,6 +1,8 @@
 import warnings
+from typing import Any
 
 import requests
+from fastapi import APIRouter
 
 from openhands.security.options import SecurityAnalyzers
 
@@ -8,10 +10,6 @@ with warnings.catch_warnings():
     warnings.simplefilter('ignore')
     import litellm
 
-from fastapi import (
-    APIRouter,
-)
-
 from openhands.controller.agent import Agent
 from openhands.core.config import LLMConfig
 from openhands.core.logger import openhands_logger as logger
@@ -21,7 +19,7 @@ from openhands.server.shared import config, server_config
 app = APIRouter(prefix='/api/options')
 
 
-@app.get('/models')
+@app.get('/models', response_model=list[str])
 async def get_litellm_models() -> list[str]:
     """Get all models supported by LiteLLM.
 
@@ -34,7 +32,7 @@ async def get_litellm_models() -> list[str]:
     ```
 
     Returns:
-        list: A sorted list of unique model names.
+        list[str]: A sorted list of unique model names.
     """
     litellm_model_list = litellm.model_list + list(litellm.model_cost.keys())
     litellm_model_list_without_bedrock = bedrock.remove_error_modelId(
@@ -74,8 +72,8 @@ async def get_litellm_models() -> list[str]:
     return list(sorted(set(model_list)))
 
 
-@app.get('/agents')
-async def get_agents():
+@app.get('/agents', response_model=list[str])
+async def get_agents() -> list[str]:
     """Get all agents supported by LiteLLM.
 
     To get the agents:
@@ -84,14 +82,13 @@ async def get_agents():
     ```
 
     Returns:
-        list: A sorted list of agent names.
+        list[str]: A sorted list of agent names.
     """
-    agents = sorted(Agent.list_agents())
-    return agents
+    return sorted(Agent.list_agents())
 
 
-@app.get('/security-analyzers')
-async def get_security_analyzers():
+@app.get('/security-analyzers', response_model=list[str])
+async def get_security_analyzers() -> list[str]:
     """Get all supported security analyzers.
 
     To get the security analyzers:
@@ -100,15 +97,16 @@ async def get_security_analyzers():
     ```
 
     Returns:
-        list: A sorted list of security analyzer names.
+        list[str]: A sorted list of security analyzer names.
     """
     return sorted(SecurityAnalyzers.keys())
 
 
-@app.get('/config')
-async def get_config():
-    """
-    Get current config
-    """
+@app.get('/config', response_model=dict[str, Any])
+async def get_config() -> dict[str, Any]:
+    """Get current config.
 
+    Returns:
+        dict[str, Any]: The current server configuration.
+    """
     return server_config.get_config()

openhands/server/routes/security.py

@@ -2,6 +2,7 @@ from fastapi import (
     APIRouter,
     HTTPException,
     Request,
+    Response,
     status,
 )
 
@@ -9,7 +10,7 @@ app = APIRouter(prefix='/api/conversations/{conversation_id}')
 
 
 @app.route('/security/{path:path}', methods=['GET', 'POST', 'PUT', 'DELETE'])
-async def security_api(request: Request):
+async def security_api(request: Request) -> Response:
     """Catch-all route for security analyzer API requests.
 
     Each request is handled directly to the security analyzer.
@@ -18,7 +19,7 @@ async def security_api(request: Request):
         request (Request): The incoming FastAPI request object.
 
     Returns:
-        Any: The response from the security analyzer.
+        Response: The response from the security analyzer.
 
     Raises:
         HTTPException: If the security analyzer is not initialized.

openhands/server/routes/settings.py

@@ -11,8 +11,8 @@ from openhands.server.shared import SettingsStoreImpl, config
 app = APIRouter(prefix='/api')
 
 
-@app.get('/settings')
-async def load_settings(request: Request) -> GETSettingsModel | None:
+@app.get('/settings', response_model=GETSettingsModel)
+async def load_settings(request: Request) -> GETSettingsModel | JSONResponse:
     try:
         user_id = get_user_id(request)
         settings_store = await SettingsStoreImpl.get_instance(config, user_id)
@@ -40,13 +40,12 @@ async def load_settings(request: Request) -> GETSettingsModel | None:
         )
 
 
-@app.post('/settings')
+@app.post('/settings', response_model=dict[str, str])
 async def store_settings(
     request: Request,
     settings: POSTSettingsModel,
 ) -> JSONResponse:
     # Check if token is valid
-
     if settings.github_token:
         try:
             # We check if the token is valid by getting the user

openhands/server/routes/trajectory.py

@@ -9,7 +9,7 @@ app = APIRouter(prefix='/api/conversations/{conversation_id}')
 
 
 @app.get('/trajectory')
-async def get_trajectory(request: Request):
+async def get_trajectory(request: Request) -> JSONResponse:
     """Get trajectory.
 
     This function retrieves the current trajectory and returns it.

openhands/server/session/agent_session.py

@@ -1,6 +1,6 @@
 import asyncio
 import time
-from typing import Callable, Optional
+from typing import Callable
 
 from pydantic import SecretStr
 
@@ -52,7 +52,7 @@ class AgentSession:
         sid: str,
         file_store: FileStore,
         monitoring_listener: MonitoringListener,
-        status_callback: Optional[Callable] = None,
+        status_callback: Callable | None = None,
         github_user_id: str | None = None,
     ):
         """Initializes a new instance of the Session class

openhands/storage/google_cloud.py

@@ -1,5 +1,5 @@
 import os
-from typing import List, Optional
+from typing import List
 
 from google.api_core.exceptions import NotFound
 from google.cloud import storage
@@ -8,7 +8,7 @@ from openhands.storage.files import FileStore
 
 
 class GoogleCloudFileStore(FileStore):
-    def __init__(self, bucket_name: Optional[str] = None) -> None:
+    def __init__(self, bucket_name: str | None = None) -> None:
         """
         Create a new FileStore. If GOOGLE_APPLICATION_CREDENTIALS is defined in the
         environment it will be used for authentication. Otherwise access will be

tests/unit/events/test_stream.py

@@ -0,0 +1,120 @@
+import pytest
+from unittest.mock import MagicMock, patch
+
+from openhands.events.event import Event
+from openhands.events.stream import EventStream
+from openhands.llm.metrics import Metrics
+
+
+class TestEventStream:
+    def test_get_metrics_empty_stream(self):
+        """Test that get_metrics returns None for an empty stream."""
+        sid = "test-stream-id"
+        file_store = MagicMock()
+        stream = EventStream(sid=sid, file_store=file_store)
+        assert stream.get_metrics() is None
+
+    def test_get_metrics_no_metrics_in_events(self):
+        """Test that get_metrics returns None when no events have metrics."""
+        sid = "test-stream-id"
+        file_store = MagicMock()
+        stream = EventStream(sid=sid, file_store=file_store)
+        event = MagicMock(spec=Event)
+        event.llm_metrics = None
+        
+        with patch.object(stream, 'get_events', return_value=[event]):
+            assert stream.get_metrics() is None
+
+    def test_get_metrics_with_metrics(self):
+        """Test that get_metrics correctly aggregates metrics from events."""
+        sid = "test-stream-id"
+        file_store = MagicMock()
+        stream = EventStream(sid=sid, file_store=file_store)
+        
+        # Create mock events with metrics
+        event1 = MagicMock(spec=Event)
+        metrics1 = Metrics(model_name="gpt-4")
+        metrics1.add_token_usage(
+            prompt_tokens=10,
+            completion_tokens=20,
+            cache_read_tokens=0,
+            cache_write_tokens=0,
+            response_id="resp1"
+        )
+        event1.llm_metrics = metrics1
+        
+        event2 = MagicMock(spec=Event)
+        metrics2 = Metrics(model_name="gpt-4")
+        metrics2.add_token_usage(
+            prompt_tokens=15,
+            completion_tokens=25,
+            cache_read_tokens=0,
+            cache_write_tokens=0,
+            response_id="resp2"
+        )
+        event2.llm_metrics = metrics2
+        
+        with patch.object(stream, 'get_events', return_value=[event1, event2]):
+            result = stream.get_metrics()
+            
+            assert result is not None
+            assert result.model_name == "gpt-4"
+            # Check token usages are merged correctly
+            total_prompt_tokens = sum(usage.prompt_tokens for usage in result.token_usages)
+            total_completion_tokens = sum(usage.completion_tokens for usage in result.token_usages)
+            assert total_prompt_tokens == 25  # 10 + 15
+            assert total_completion_tokens == 45  # 20 + 25
+            assert len(result.token_usages) == 2
+
+    def test_get_metrics_with_exception(self):
+        """Test that get_metrics handles exceptions gracefully."""
+        sid = "test-stream-id"
+        file_store = MagicMock()
+        stream = EventStream(sid=sid, file_store=file_store)
+        
+        with patch.object(stream, 'get_events', side_effect=Exception("Test exception")):
+            assert stream.get_metrics() is None
+
+    def test_get_metrics_with_mixed_events(self):
+        """Test that get_metrics correctly handles a mix of events with and without metrics."""
+        sid = "test-stream-id"
+        file_store = MagicMock()
+        stream = EventStream(sid=sid, file_store=file_store)
+        
+        # Create mock events, some with metrics and some without
+        event1 = MagicMock(spec=Event)
+        metrics1 = Metrics(model_name="gpt-4")
+        metrics1.add_token_usage(
+            prompt_tokens=10,
+            completion_tokens=20,
+            cache_read_tokens=0,
+            cache_write_tokens=0,
+            response_id="resp1"
+        )
+        event1.llm_metrics = metrics1
+        
+        event2 = MagicMock(spec=Event)
+        event2.llm_metrics = None
+        
+        event3 = MagicMock(spec=Event)
+        metrics3 = Metrics(model_name="gpt-4")
+        metrics3.add_token_usage(
+            prompt_tokens=15,
+            completion_tokens=25,
+            cache_read_tokens=0,
+            cache_write_tokens=0,
+            response_id="resp3"
+        )
+        event3.llm_metrics = metrics3
+        
+        with patch.object(stream, 'get_events', return_value=[event1, event2, event3]):
+            result = stream.get_metrics()
+            
+            assert result is not None
+            assert result.model_name == "gpt-4"
+            # Check token usages are merged correctly
+            total_prompt_tokens = sum(usage.prompt_tokens for usage in result.token_usages)
+            total_completion_tokens = sum(usage.completion_tokens for usage in result.token_usages)
+            assert total_prompt_tokens == 25  # 10 + 15
+            assert total_completion_tokens == 45  # 20 + 25
+            assert len(result.token_usages) == 2

tests/unit/server/routes/test_conversation_metrics.py

@@ -0,0 +1,139 @@
+import pytest
+from unittest.mock import MagicMock, patch
+from fastapi import Request, status
+from fastapi.responses import JSONResponse
+
+from openhands.server.routes.conversation import get_conversation_metrics
+from openhands.llm.metrics import Metrics, TokenUsage
+
+
+@pytest.fixture
+def mock_request():
+    """Create a mock request with a conversation."""
+    request = MagicMock(spec=Request)
+    request.state.conversation = MagicMock()
+    request.state.conversation.runtime = MagicMock()
+    request.state.conversation.event_stream = MagicMock()
+    return request
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_success(mock_request):
+    """Test successful retrieval of conversation metrics."""
+    # Setup mock metrics
+    metrics = Metrics()
+    metrics.token_usages = [
+        TokenUsage(
+            prompt_tokens=100, 
+            completion_tokens=50, 
+            model="test-model", 
+            cache_read_tokens=0, 
+            cache_write_tokens=0, 
+            response_id="test-response-1"
+        ),
+        TokenUsage(
+            prompt_tokens=200, 
+            completion_tokens=150, 
+            model="test-model", 
+            cache_read_tokens=0, 
+            cache_write_tokens=0, 
+            response_id="test-response-2"
+        ),
+    ]
+    metrics.accumulated_cost = 0.25
+    
+    # Configure mock to return metrics
+    mock_request.state.conversation.event_stream.get_metrics.return_value = metrics
+    
+    # Call the endpoint
+    response = await get_conversation_metrics(mock_request)
+    
+    # Verify response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_200_OK
+    
+    # Extract content from JSONResponse
+    content = response.body.decode('utf-8')
+    import json
+    content_dict = json.loads(content)
+    
+    # Verify metrics
+    assert content_dict['accumulated_cost'] == 0.25
+    assert content_dict['total_prompt_tokens'] == 300
+    assert content_dict['total_completion_tokens'] == 200
+    assert content_dict['total_tokens'] == 500
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_no_metrics(mock_request):
+    """Test handling when no metrics are available."""
+    # Configure mock to return None for metrics
+    mock_request.state.conversation.event_stream.get_metrics.return_value = None
+    
+    # Call the endpoint
+    response = await get_conversation_metrics(mock_request)
+    
+    # Verify response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_200_OK
+    
+    # Extract content from JSONResponse
+    content = response.body.decode('utf-8')
+    import json
+    content_dict = json.loads(content)
+    
+    # Verify default metrics
+    assert content_dict['accumulated_cost'] == 0.0
+    assert content_dict['total_prompt_tokens'] == 0
+    assert content_dict['total_completion_tokens'] == 0
+    assert content_dict['total_tokens'] == 0
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_no_conversation():
+    """Test handling when no conversation is found in request state."""
+    # Create a request without conversation
+    request = MagicMock(spec=Request)
+    request.state = MagicMock()
+    
+    # Remove conversation attribute
+    delattr(request.state, 'conversation')
+    
+    # Call the endpoint
+    response = await get_conversation_metrics(request)
+    
+    # Verify response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_500_INTERNAL_SERVER_ERROR
+    
+    # Extract content from JSONResponse
+    content = response.body.decode('utf-8')
+    import json
+    content_dict = json.loads(content)
+    
+    # Verify error message
+    assert 'error' in content_dict
+    assert 'No conversation found in request state' in content_dict['error']
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_exception(mock_request):
+    """Test handling when an exception occurs."""
+    # Configure mock to raise an exception
+    mock_request.state.conversation.event_stream.get_metrics.side_effect = Exception("Test exception")
+    
+    # Call the endpoint
+    response = await get_conversation_metrics(mock_request)
+    
+    # Verify response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_500_INTERNAL_SERVER_ERROR
+    
+    # Extract content from JSONResponse
+    content = response.body.decode('utf-8')
+    import json
+    content_dict = json.loads(content)
+    
+    # Verify error message
+    assert 'error' in content_dict
+    assert 'Test exception' in content_dict['error']

tests/unit/test_config.py

@@ -16,7 +16,9 @@ from openhands.core.config import (
     load_from_toml,
 )
 from openhands.core.config.condenser_config import (
+    LLMSummarizingCondenserConfig,
     NoOpCondenserConfig,
+    RecentEventsCondenserConfig,
 )
 from openhands.core.logger import openhands_logger
 
@@ -566,13 +568,243 @@ def test_cache_dir_creation(default_config, tmpdir):
     assert os.path.exists(default_config.cache_dir)
 
 
-def test_agent_config_condenser_default():
-    """Test that default agent condenser is NoOpCondenser."""
-    config = AppConfig()
+def test_agent_config_condenser_with_no_enabled():
+    """Test default agent condenser with enable_default_condenser=False."""
+    config = AppConfig(enable_default_condenser=False)
     agent_config = config.get_agent_config()
     assert isinstance(agent_config.condenser, NoOpCondenserConfig)
 
 
+def test_condenser_config_from_toml_basic(default_config, temp_toml_file):
+    """Test loading basic condenser configuration from TOML."""
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[condenser]
+type = "recent"
+keep_first = 3
+max_events = 15
+""")
+
+    load_from_toml(default_config, temp_toml_file)
+
+    # Verify that the condenser config is correctly assigned to the default agent config
+    agent_config = default_config.get_agent_config()
+    assert isinstance(agent_config.condenser, RecentEventsCondenserConfig)
+    assert agent_config.condenser.keep_first == 3
+    assert agent_config.condenser.max_events == 15
+
+    # We can also verify the function works directly
+    from openhands.core.config.condenser_config import (
+        condenser_config_from_toml_section,
+    )
+
+    condenser_data = {'type': 'recent', 'keep_first': 3, 'max_events': 15}
+    condenser_mapping = condenser_config_from_toml_section(condenser_data)
+
+    assert 'condenser' in condenser_mapping
+    assert isinstance(condenser_mapping['condenser'], RecentEventsCondenserConfig)
+    assert condenser_mapping['condenser'].keep_first == 3
+    assert condenser_mapping['condenser'].max_events == 15
+
+
+def test_condenser_config_from_toml_with_llm_reference(default_config, temp_toml_file):
+    """Test loading condenser configuration with LLM reference from TOML."""
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[llm.condenser_llm]
+model = "gpt-4"
+api_key = "test-key"
+
+[condenser]
+type = "llm"
+llm_config = "condenser_llm"
+keep_first = 2
+max_size = 50
+""")
+
+    load_from_toml(default_config, temp_toml_file)
+
+    # Verify that the LLM config was loaded
+    assert 'condenser_llm' in default_config.llms
+    assert default_config.llms['condenser_llm'].model == 'gpt-4'
+
+    # Verify that the condenser config is correctly assigned to the default agent config
+    agent_config = default_config.get_agent_config()
+    assert isinstance(agent_config.condenser, LLMSummarizingCondenserConfig)
+    assert agent_config.condenser.keep_first == 2
+    assert agent_config.condenser.max_size == 50
+    assert agent_config.condenser.llm_config.model == 'gpt-4'
+
+    # Test the condenser config with the LLM reference
+    from openhands.core.config.condenser_config import (
+        condenser_config_from_toml_section,
+    )
+
+    condenser_data = {
+        'type': 'llm',
+        'llm_config': 'condenser_llm',
+        'keep_first': 2,
+        'max_size': 50,
+    }
+    condenser_mapping = condenser_config_from_toml_section(
+        condenser_data, default_config.llms
+    )
+
+    assert 'condenser' in condenser_mapping
+    assert isinstance(condenser_mapping['condenser'], LLMSummarizingCondenserConfig)
+    assert condenser_mapping['condenser'].keep_first == 2
+    assert condenser_mapping['condenser'].max_size == 50
+    assert condenser_mapping['condenser'].llm_config.model == 'gpt-4'
+
+
+def test_condenser_config_from_toml_with_missing_llm_reference(
+    default_config, temp_toml_file
+):
+    """Test loading condenser configuration with missing LLM reference from TOML."""
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[condenser]
+type = "llm"
+llm_config = "missing_llm"
+keep_first = 2
+max_size = 50
+""")
+
+    load_from_toml(default_config, temp_toml_file)
+
+    # Test the condenser config with a missing LLM reference
+    from openhands.core.config.condenser_config import (
+        condenser_config_from_toml_section,
+    )
+
+    condenser_data = {
+        'type': 'llm',
+        'llm_config': 'missing_llm',
+        'keep_first': 2,
+        'max_size': 50,
+    }
+    condenser_mapping = condenser_config_from_toml_section(
+        condenser_data, default_config.llms
+    )
+
+    assert 'condenser' in condenser_mapping
+    assert isinstance(condenser_mapping['condenser'], NoOpCondenserConfig)
+    # Should not have a default LLMConfig when the reference is missing
+    assert not hasattr(condenser_mapping['condenser'], 'llm_config')
+
+
+def test_condenser_config_from_toml_with_invalid_config(default_config, temp_toml_file):
+    """Test loading invalid condenser configuration from TOML."""
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[condenser]
+type = "invalid_type"
+""")
+
+    load_from_toml(default_config, temp_toml_file)
+
+    # Test the condenser config with an invalid type
+    from openhands.core.config.condenser_config import (
+        condenser_config_from_toml_section,
+    )
+
+    condenser_data = {'type': 'invalid_type'}
+    condenser_mapping = condenser_config_from_toml_section(condenser_data)
+
+    # Should default to NoOpCondenserConfig when the type is invalid
+    assert 'condenser' in condenser_mapping
+    assert isinstance(condenser_mapping['condenser'], NoOpCondenserConfig)
+
+
+def test_condenser_config_from_toml_with_validation_error(
+    default_config, temp_toml_file
+):
+    """Test loading condenser configuration with validation error from TOML."""
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[condenser]
+type = "recent"
+keep_first = -1  # Invalid: must be >= 0
+max_events = 0   # Invalid: must be >= 1
+""")
+
+    load_from_toml(default_config, temp_toml_file)
+
+    # Test the condenser config with validation errors
+    from openhands.core.config.condenser_config import (
+        condenser_config_from_toml_section,
+    )
+
+    condenser_data = {'type': 'recent', 'keep_first': -1, 'max_events': 0}
+    condenser_mapping = condenser_config_from_toml_section(condenser_data)
+
+    # Should default to NoOpCondenserConfig when validation fails
+    assert 'condenser' in condenser_mapping
+    assert isinstance(condenser_mapping['condenser'], NoOpCondenserConfig)
+
+
+def test_default_condenser_behavior_enabled(default_config, temp_toml_file):
+    """Test the default condenser behavior when enable_default_condenser is True."""
+    # Create a minimal TOML file with no condenser section
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[core]
+# Empty core section, no condenser section
+""")
+
+    # Set enable_default_condenser to True
+    default_config.enable_default_condenser = True
+    load_from_toml(default_config, temp_toml_file)
+
+    # Verify the default agent config has LLMSummarizingCondenserConfig
+    agent_config = default_config.get_agent_config()
+    assert isinstance(agent_config.condenser, LLMSummarizingCondenserConfig)
+    assert agent_config.condenser.keep_first == 1
+    assert agent_config.condenser.max_size == 100
+
+
+def test_default_condenser_behavior_disabled(default_config, temp_toml_file):
+    """Test the default condenser behavior when enable_default_condenser is False."""
+    # Create a minimal TOML file with no condenser section
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[core]
+# Empty core section, no condenser section
+""")
+
+    # Set enable_default_condenser to False
+    default_config.enable_default_condenser = False
+    load_from_toml(default_config, temp_toml_file)
+
+    # Verify the agent config uses NoOpCondenserConfig
+    agent_config = default_config.get_agent_config()
+    assert isinstance(agent_config.condenser, NoOpCondenserConfig)
+
+
+def test_default_condenser_explicit_toml_override(default_config, temp_toml_file):
+    """Test that explicit condenser in TOML takes precedence over the default."""
+    # Set enable_default_condenser to True
+    default_config.enable_default_condenser = True
+
+    # Create a TOML file with an explicit condenser section
+    with open(temp_toml_file, 'w', encoding='utf-8') as toml_file:
+        toml_file.write("""
+[condenser]
+type = "recent"
+keep_first = 3
+max_events = 15
+""")
+
+    # Load the config
+    load_from_toml(default_config, temp_toml_file)
+
+    # Verify the explicit condenser from TOML takes precedence
+    agent_config = default_config.get_agent_config()
+    assert isinstance(agent_config.condenser, RecentEventsCondenserConfig)
+    assert agent_config.condenser.keep_first == 3
+    assert agent_config.condenser.max_events == 15
+
+
 def test_api_keys_repr_str():
     # Test LLMConfig
     llm_config = LLMConfig(

tests/unit/test_conversation_metrics_api.py

@@ -0,0 +1,164 @@
+import pytest
+from unittest.mock import MagicMock, patch
+
+from fastapi import Request, status
+from fastapi.responses import JSONResponse
+from openhands.llm.metrics import Metrics, TokenUsage
+from openhands.server.routes.conversation import get_conversation_metrics
+
+
+@pytest.fixture
+def mock_metrics():
+    metrics = Metrics()
+    metrics.accumulated_cost = 0.25
+    metrics.token_usages = [
+        TokenUsage(
+            model="gpt-4", 
+            prompt_tokens=100, 
+            completion_tokens=50, 
+            cache_read_tokens=0, 
+            cache_write_tokens=0, 
+            response_id="resp1"
+        ),
+        TokenUsage(
+            model="gpt-4", 
+            prompt_tokens=200, 
+            completion_tokens=75, 
+            cache_read_tokens=0, 
+            cache_write_tokens=0, 
+            response_id="resp2"
+        ),
+    ]
+    return metrics
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_success(mock_metrics):
+    """Test that the metrics endpoint returns the correct metrics data."""
+    
+    # Create a mock request with a conversation that has metrics
+    mock_event_stream = MagicMock()
+    mock_event_stream.get_metrics.return_value = mock_metrics
+    
+    mock_conversation = MagicMock()
+    mock_conversation.event_stream = mock_event_stream
+    
+    mock_request = MagicMock()
+    mock_request.state.conversation = mock_conversation
+    
+    # Call the endpoint function directly
+    response = await get_conversation_metrics(mock_request)
+    
+    # Check the response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_200_OK
+    
+    # Extract the content from the response
+    content = response.body.decode('utf-8')
+    import json
+    data = json.loads(content)
+    
+    # Verify the metrics data
+    assert data["accumulated_cost"] == 0.25
+    assert data["total_prompt_tokens"] == 300
+    assert data["total_completion_tokens"] == 125
+    assert data["total_tokens"] == 425
+    
+    # Verify the get_metrics method was called
+    mock_conversation.event_stream.get_metrics.assert_called_once()
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_no_metrics():
+    """Test that the metrics endpoint handles the case where no metrics are available."""
+    
+    # Create a mock request with a conversation that has no metrics
+    mock_event_stream = MagicMock()
+    mock_event_stream.get_metrics.return_value = None
+    
+    mock_conversation = MagicMock()
+    mock_conversation.event_stream = mock_event_stream
+    
+    mock_request = MagicMock()
+    mock_request.state.conversation = mock_conversation
+    
+    # Call the endpoint function directly
+    response = await get_conversation_metrics(mock_request)
+    
+    # Check the response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_200_OK
+    
+    # Extract the content from the response
+    content = response.body.decode('utf-8')
+    import json
+    data = json.loads(content)
+    
+    # Verify the metrics data
+    assert data["accumulated_cost"] == 0.0
+    assert data["total_prompt_tokens"] == 0
+    assert data["total_completion_tokens"] == 0
+    assert data["total_tokens"] == 0
+    
+    # Verify the get_metrics method was called
+    mock_conversation.event_stream.get_metrics.assert_called_once()
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_no_conversation():
+    """Test that the metrics endpoint handles the case where no conversation is found."""
+    
+    # Create a mock request with no conversation attribute
+    mock_request = MagicMock()
+    mock_request.state = MagicMock()
+    # Intentionally not setting request.state.conversation
+    
+    # Call the endpoint function directly
+    response = await get_conversation_metrics(mock_request)
+    
+    # Check the response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_500_INTERNAL_SERVER_ERROR
+    
+    # Extract the content from the response
+    content = response.body.decode('utf-8')
+    import json
+    data = json.loads(content)
+    
+    # Verify the error message
+    assert "error" in data
+    assert "No conversation found" in data["error"] or "conversation" in data["error"].lower()
+
+
+@pytest.mark.asyncio
+async def test_get_conversation_metrics_exception():
+    """Test that the metrics endpoint handles exceptions gracefully."""
+    
+    # Create a mock request with a conversation that raises an exception
+    mock_event_stream = MagicMock()
+    mock_event_stream.get_metrics.side_effect = Exception("Test exception")
+    
+    mock_conversation = MagicMock()
+    mock_conversation.event_stream = mock_event_stream
+    
+    mock_request = MagicMock()
+    mock_request.state.conversation = mock_conversation
+    
+    # Call the endpoint function directly
+    response = await get_conversation_metrics(mock_request)
+    
+    # Check the response
+    assert isinstance(response, JSONResponse)
+    assert response.status_code == status.HTTP_500_INTERNAL_SERVER_ERROR
+    
+    # Extract the content from the response
+    content = response.body.decode('utf-8')
+    import json
+    data = json.loads(content)
+    
+    # Verify the error message
+    assert "error" in data
+    assert "Test exception" in data["error"]
+    
+    # Verify the get_metrics method was called
+    mock_conversation.event_stream.get_metrics.assert_called_once()