feat(llm): Add Claude 3.7 backend configurations (#6937)

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

.github/pull_request_template.md

@@ -1,11 +1,12 @@
-**End-user friendly description of the problem this fixes or functionality that this introduces**
-
-- [ ] Include this change in the Release Notes. If checked, you must provide an **end-user friendly** description for your change below
-
----
-**Give a summary of what the PR does, explaining any non-trivial design decisions**
+- [ ] This change is worth documenting at https://docs.all-hands.dev/
+- [ ] Include this change in the Release Notes. If checked, you **must** provide an **end-user friendly** description for your change below
 
+**End-user friendly description of the problem this fixes or functionality that this introduces.**
 
 
 ---
-**Link of any specific issues this addresses**
+**Give a summary of what the PR does, explaining any non-trivial design decisions.**
+
+
+---
+**Link of any specific issues this addresses.**

.github/workflows/py-unit-tests.yml

@@ -48,7 +48,7 @@ jobs:
       - name: Build Environment
         run: make build
       - name: Run Tests
-        run: poetry run pytest --forked -n auto --cov=openhands --cov-report=xml -svv ./tests/unit --ignore=tests/unit/test_memory.py
+        run: poetry run pytest --forked -n auto --cov=openhands --cov-report=xml -svv ./tests/unit --ignore=tests/unit/test_long_term_memory.py
       - name: Upload coverage to Codecov
         uses: codecov/codecov-action@v5
         env:

.gitignore

@@ -233,6 +233,3 @@ containers/runtime/Dockerfile
 containers/runtime/project.tar.gz
 containers/runtime/code
 **/node_modules/
-
-# regression test workspaces
-tests/regression/cases/*/workspace/

Development.md

@@ -100,7 +100,7 @@ poetry run pytest ./tests/unit/test_*.py
 To reduce build time (e.g., if no changes were made to the client-runtime component), you can use an existing Docker container image by
 setting the SANDBOX_RUNTIME_CONTAINER_IMAGE environment variable to the desired Docker image.
 
-Example: `export SANDBOX_RUNTIME_CONTAINER_IMAGE=ghcr.io/all-hands-ai/runtime:0.25-nikolaik`
+Example: `export SANDBOX_RUNTIME_CONTAINER_IMAGE=ghcr.io/all-hands-ai/runtime:0.26-nikolaik`
 
 ## Develop inside Docker container
 

Makefile

@@ -1,4 +1,4 @@
-SHELL=/bin/bash
+SHELL=/usr/bin/env bash
 # Makefile for OpenHands project
 
 # Variables

README.md

@@ -43,17 +43,17 @@ See the [Running OpenHands](https://docs.all-hands.dev/modules/usage/installatio
 system requirements and more information.
 
 ```bash
-docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e LOG_ALL_EVENTS=true \
     -v /var/run/docker.sock:/var/run/docker.sock \
     -v ~/.openhands-state:/.openhands-state \
     -p 3000:3000 \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25
+    docker.all-hands.dev/all-hands-ai/openhands:0.26
 ```
 
 You'll find OpenHands running at [http://localhost:3000](http://localhost:3000)!

build.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -e
 
 poetry build -v

config.template.toml

@@ -17,6 +17,12 @@
 #modal_api_token_id = ""
 #modal_api_token_secret = ""
 
+# API key for Daytona
+#daytona_api_key = ""
+
+# Daytona Target
+#daytona_target = ""
+
 # Base path for the workspace
 workspace_base = "./workspace"
 

containers/build.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 # Initialize variables with default values

containers/dev/compose.yml

@@ -11,7 +11,7 @@ services:
       - BACKEND_HOST=${BACKEND_HOST:-"0.0.0.0"}
       - SANDBOX_API_HOSTNAME=host.docker.internal
       #
-      - SANDBOX_RUNTIME_CONTAINER_IMAGE=${SANDBOX_RUNTIME_CONTAINER_IMAGE:-ghcr.io/all-hands-ai/runtime:0.25-nikolaik}
+      - SANDBOX_RUNTIME_CONTAINER_IMAGE=${SANDBOX_RUNTIME_CONTAINER_IMAGE:-ghcr.io/all-hands-ai/runtime:0.26-nikolaik}
       - SANDBOX_USER_ID=${SANDBOX_USER_ID:-1234}
       - WORKSPACE_MOUNT_PATH=${WORKSPACE_BASE:-$PWD/workspace}
     ports:

containers/dev/dev.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -o pipefail
 
 function get_docker() {

docker-compose.yml

@@ -7,7 +7,7 @@ services:
     image: openhands:latest
     container_name: openhands-app-${DATE:-}
     environment:
-      - SANDBOX_RUNTIME_CONTAINER_IMAGE=${SANDBOX_RUNTIME_CONTAINER_IMAGE:-docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik}
+      - SANDBOX_RUNTIME_CONTAINER_IMAGE=${SANDBOX_RUNTIME_CONTAINER_IMAGE:-docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik}
       #- SANDBOX_USER_ID=${SANDBOX_USER_ID:-1234} # enable this only if you want a specific non-root sandbox user but you will have to manually adjust permissions of openhands-state for this user
       - WORKSPACE_MOUNT_PATH=${WORKSPACE_BASE:-$PWD/workspace}
     ports:

docs/i18n/fr/docusaurus-plugin-content-docs/current/usage/how-to/cli-mode.md

@@ -52,7 +52,7 @@ LLM_API_KEY="sk_test_12345"
 ```bash
 docker run -it \
     --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e SANDBOX_USER_ID=$(id -u) \
     -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
     -e LLM_API_KEY=$LLM_API_KEY \
@@ -61,7 +61,7 @@ docker run -it \
     -v /var/run/docker.sock:/var/run/docker.sock \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app-$(date +%Y%m%d%H%M%S) \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.26 \
     python -m openhands.core.cli
 ```
 

docs/i18n/fr/docusaurus-plugin-content-docs/current/usage/how-to/headless-mode.md

@@ -46,7 +46,7 @@ LLM_API_KEY="sk_test_12345"
 ```bash
 docker run -it \
     --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e SANDBOX_USER_ID=$(id -u) \
     -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
     -e LLM_API_KEY=$LLM_API_KEY \
@@ -56,6 +56,6 @@ docker run -it \
     -v /var/run/docker.sock:/var/run/docker.sock \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app-$(date +%Y%m%d%H%M%S) \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.26 \
     python -m openhands.core.main -t "write a bash script that prints hi" --no-auto-continue
 ```

docs/i18n/fr/docusaurus-plugin-content-docs/current/usage/installation.mdx

@@ -13,16 +13,16 @@
 La façon la plus simple d'exécuter OpenHands est avec Docker.
 
 ```bash
-docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e LOG_ALL_EVENTS=true \
     -v /var/run/docker.sock:/var/run/docker.sock \
     -p 3000:3000 \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25
+    docker.all-hands.dev/all-hands-ai/openhands:0.26
 ```
 
 Vous pouvez également exécuter OpenHands en mode [headless scriptable](https://docs.all-hands.dev/modules/usage/how-to/headless-mode), en tant que [CLI interactive](https://docs.all-hands.dev/modules/usage/how-to/cli-mode), ou en utilisant l'[Action GitHub OpenHands](https://docs.all-hands.dev/modules/usage/how-to/github-action).

docs/i18n/fr/docusaurus-plugin-content-docs/current/usage/runtimes.md

@@ -13,7 +13,7 @@ C'est le Runtime par défaut qui est utilisé lorsque vous démarrez OpenHands.
 
 ```
 docker run # ...
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -v /var/run/docker.sock:/var/run/docker.sock \
     # ...
 ```

docs/i18n/zh-Hans/docusaurus-plugin-content-docs/current/usage/how-to/cli-mode.md

@@ -50,7 +50,7 @@ LLM_API_KEY="sk_test_12345"
 ```bash
 docker run -it \
     --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e SANDBOX_USER_ID=$(id -u) \
     -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
     -e LLM_API_KEY=$LLM_API_KEY \
@@ -59,7 +59,7 @@ docker run -it \
     -v /var/run/docker.sock:/var/run/docker.sock \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app-$(date +%Y%m%d%H%M%S) \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.26 \
     python -m openhands.core.cli
 ```
 

docs/i18n/zh-Hans/docusaurus-plugin-content-docs/current/usage/how-to/headless-mode.md

@@ -47,7 +47,7 @@ LLM_API_KEY="sk_test_12345"
 ```bash
 docker run -it \
     --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e SANDBOX_USER_ID=$(id -u) \
     -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
     -e LLM_API_KEY=$LLM_API_KEY \
@@ -57,6 +57,6 @@ docker run -it \
     -v /var/run/docker.sock:/var/run/docker.sock \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app-$(date +%Y%m%d%H%M%S) \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.26 \
     python -m openhands.core.main -t "write a bash script that prints hi" --no-auto-continue
 ```

docs/i18n/zh-Hans/docusaurus-plugin-content-docs/current/usage/installation.mdx

@@ -11,16 +11,16 @@
 在 Docker 中运行 OpenHands 是最简单的方式。
 
 ```bash
-docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e LOG_ALL_EVENTS=true \
     -v /var/run/docker.sock:/var/run/docker.sock \
     -p 3000:3000 \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25
+    docker.all-hands.dev/all-hands-ai/openhands:0.26
 ```
 
 你也可以在可脚本化的[无头模式](https://docs.all-hands.dev/modules/usage/how-to/headless-mode)下运行 OpenHands，作为[交互式 CLI](https://docs.all-hands.dev/modules/usage/how-to/cli-mode)，或使用 [OpenHands GitHub Action](https://docs.all-hands.dev/modules/usage/how-to/github-action)。

docs/i18n/zh-Hans/docusaurus-plugin-content-docs/current/usage/runtimes.md

@@ -11,7 +11,7 @@
 
 ```
 docker run # ...
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -v /var/run/docker.sock:/var/run/docker.sock \
     # ...
 ```

docs/modules/usage/how-to/cli-mode.md

@@ -35,7 +35,7 @@ To run OpenHands in CLI mode with Docker:
 ```bash
 docker run -it \
     --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e SANDBOX_USER_ID=$(id -u) \
     -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
     -e LLM_API_KEY=$LLM_API_KEY \
@@ -45,7 +45,7 @@ docker run -it \
     -v ~/.openhands-state:/.openhands-state \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app-$(date +%Y%m%d%H%M%S) \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.26 \
     python -m openhands.core.cli
 ```
 

docs/modules/usage/how-to/headless-mode.md

@@ -32,7 +32,7 @@ To run OpenHands in Headless mode with Docker:
 ```bash
 docker run -it \
     --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e SANDBOX_USER_ID=$(id -u) \
     -e WORKSPACE_MOUNT_PATH=$WORKSPACE_BASE \
     -e LLM_API_KEY=$LLM_API_KEY \
@@ -43,7 +43,7 @@ docker run -it \
     -v ~/.openhands-state:/.openhands-state \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app-$(date +%Y%m%d%H%M%S) \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.26 \
     python -m openhands.core.main -t "write a bash script that prints hi"
 ```
 

docs/modules/usage/installation.mdx

@@ -58,17 +58,17 @@ A system with a modern processor and a minimum of **4GB RAM** is recommended to
 The easiest way to run OpenHands is in Docker.
 
 ```bash
-docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -e LOG_ALL_EVENTS=true \
     -v /var/run/docker.sock:/var/run/docker.sock \
     -v ~/.openhands-state:/.openhands-state \
     -p 3000:3000 \
     --add-host host.docker.internal:host-gateway \
     --name openhands-app \
-    docker.all-hands.dev/all-hands-ai/openhands:0.25
+    docker.all-hands.dev/all-hands-ai/openhands:0.26
 ```
 
 You'll find OpenHands running at http://localhost:3000!

docs/modules/usage/runtimes.md

@@ -16,7 +16,7 @@ some flags being passed to `docker run` that make this possible:
 
 ```
 docker run # ...
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.26-nikolaik \
     -v /var/run/docker.sock:/var/run/docker.sock \
     # ...
 ```

evaluation/benchmarks/EDA/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/agent_bench/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/aider_bench/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/biocoder/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/bird/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/browsing_delegation/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/commit0_bench/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/discoverybench/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/gaia/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/gorilla/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/gpqa/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/humanevalfix/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/logic_reasoning/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/miniwob/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/mint/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/ml_bench/scripts/cleanup.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # Step 1: Stop all running containers
 echo "Stopping all running containers..."

evaluation/benchmarks/ml_bench/scripts/run_analysis.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 RESULT_FILE=$1
 MODEL_CONFIG=$2

evaluation/benchmarks/ml_bench/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/scienceagentbench/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/swe_bench/scripts/docker/pull_all_eval_docker.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -e
 
 LEVEL=$1

evaluation/benchmarks/swe_bench/scripts/docker/push_eval_docker.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 # This is ONLY used for pushing docker images created by https://github.com/princeton-nlp/SWE-bench/blob/main/docs/20240627_docker/README.md
 

evaluation/benchmarks/swe_bench/scripts/eval/convert_oh_folder_to_swebench_submission.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 FOLDER_PATH=$1
 NEW_FOLDER_PATH=${FOLDER_PATH}.swebench_submission

evaluation/benchmarks/swe_bench/scripts/eval_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 PROCESS_FILEPATH=$1
 if [ -z "$PROCESS_FILEPATH" ]; then

evaluation/benchmarks/swe_bench/scripts/eval_infer_remote.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 INPUT_FILE=$1

evaluation/benchmarks/swe_bench/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/swe_bench/scripts/setup/instance_swe_entry.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 source ~/.bashrc
 SWEUTIL_DIR=/swe_util

evaluation/benchmarks/swe_bench/scripts/setup/prepare_swe_utils.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 set -e
 EVAL_WORKSPACE="evaluation/benchmarks/swe_bench/eval_workspace"

evaluation/benchmarks/swe_bench/scripts/setup/swe_entry.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 set -e
 

evaluation/benchmarks/the_agent_company/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 ##################################################################################################
 # Adapted from https://github.com/TheAgentCompany/TheAgentCompany/blob/main/evaluation/run_eval.sh

evaluation/benchmarks/toolqa/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/visualwebarena/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/benchmarks/webarena/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/integration_tests/scripts/run_infer.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -eo pipefail
 
 source "evaluation/utils/version_control.sh"

evaluation/regression/cases/hello-name/start/hello_world.sh

@@ -1,2 +1,2 @@
-#!/bin/bash
+#!/usr/bin/env bash
 echo "hello world"

evaluation/utils/scripts/cleanup_remote_runtime.sh

@@ -1,4 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 
 
 # API base URL

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

@@ -65,6 +65,12 @@ describe("extractModelAndProvider", () => {
       separator: "/",
     });
 
+    expect(extractModelAndProvider("claude-3-7-sonnet-20250219")).toEqual({
+      provider: "anthropic",
+      model: "claude-3-7-sonnet-20250219",
+      separator: "/",
+    });
+
     expect(extractModelAndProvider("claude-3-haiku-20240307")).toEqual({
       provider: "anthropic",
       model: "claude-3-haiku-20240307",

frontend/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "openhands-frontend",
-  "version": "0.25.0",
+  "version": "0.26.0",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "openhands-frontend",
-      "version": "0.25.0",
+      "version": "0.26.0",
       "dependencies": {
         "@heroui/react": "2.6.14",
         "@monaco-editor/react": "^4.7.0-rc.0",

frontend/package.json

@@ -1,6 +1,6 @@
 {
   "name": "openhands-frontend",
-  "version": "0.25.0",
+  "version": "0.26.0",
   "private": true,
   "type": "module",
   "engines": {

frontend/src/components/features/settings/styled-switch-component.tsx

@@ -12,13 +12,14 @@ export function StyledSwitchComponent({
       className={cn(
         "w-12 h-6 rounded-xl flex items-center p-1.5 cursor-pointer",
         isToggled && "justify-end bg-primary",
-        !isToggled && "justify-start bg-[#1F2228] border border-tertiary-alt",
+        !isToggled &&
+          "justify-start bg-base-secondary border border-tertiary-light",
       )}
     >
       <div
         className={cn(
-          "bg-[#1F2228] w-3 h-3 rounded-xl",
-          isToggled ? "bg-[#1F2228]" : "bg-tertiary-alt",
+          "w-3 h-3 rounded-xl",
+          isToggled ? "bg-base-secondary" : "bg-tertiary-light",
         )}
       />
     </div>

frontend/src/utils/verified-models.ts

@@ -3,6 +3,7 @@ export const VERIFIED_PROVIDERS = ["openai", "azure", "anthropic", "deepseek"];
 export const VERIFIED_MODELS = [
   "o3-mini-2025-01-31",
   "claude-3-5-sonnet-20241022",
+  "claude-3-7-sonnet-20250219",
   "deepseek-chat",
 ];
 
@@ -31,4 +32,5 @@ export const VERIFIED_ANTHROPIC_MODELS = [
   "claude-3-haiku-20240307",
   "claude-3-opus-20240229",
   "claude-3-sonnet-20240229",
+  "claude-3-7-sonnet-20250219",
 ];

openhands/core/cli.py

@@ -100,6 +100,7 @@ async def main(loop: asyncio.AbstractEventLoop):
     initial_user_action = MessageAction(content=task_str) if task_str else None
 
     sid = str(uuid4())
+    display_message(f'Session ID: {sid}')
 
     runtime = create_runtime(config, sid=sid, headless_mode=True)
     await runtime.connect()

openhands/core/config/__init__.py

@@ -5,6 +5,7 @@ from openhands.core.config.config_utils import (
     OH_MAX_ITERATIONS,
     get_field_info,
 )
+from openhands.core.config.extended_config import ExtendedConfig
 from openhands.core.config.llm_config import LLMConfig
 from openhands.core.config.sandbox_config import SandboxConfig
 from openhands.core.config.security_config import SecurityConfig
@@ -28,6 +29,7 @@ __all__ = [
     'LLMConfig',
     'SandboxConfig',
     'SecurityConfig',
+    'ExtendedConfig',
     'load_app_config',
     'load_from_env',
     'load_from_toml',

openhands/core/config/app_config.py

@@ -9,6 +9,7 @@ from openhands.core.config.config_utils import (
     OH_MAX_ITERATIONS,
     model_defaults_to_dict,
 )
+from openhands.core.config.extended_config import ExtendedConfig
 from openhands.core.config.llm_config import LLMConfig
 from openhands.core.config.sandbox_config import SandboxConfig
 from openhands.core.config.security_config import SecurityConfig
@@ -52,6 +53,7 @@ class AppConfig(BaseModel):
     default_agent: str = Field(default=OH_DEFAULT_AGENT)
     sandbox: SandboxConfig = Field(default_factory=SandboxConfig)
     security: SecurityConfig = Field(default_factory=SecurityConfig)
+    extended: ExtendedConfig = Field(default_factory=lambda: ExtendedConfig({}))
     runtime: str = Field(default='docker')
     file_store: str = Field(default='local')
     file_store_path: str = Field(default='/tmp/openhands_file_store')
@@ -75,6 +77,9 @@ class AppConfig(BaseModel):
     file_uploads_restrict_file_types: bool = Field(default=False)
     file_uploads_allowed_extensions: list[str] = Field(default_factory=lambda: ['.*'])
     runloop_api_key: SecretStr | None = Field(default=None)
+    daytona_api_key: SecretStr | None = Field(default=None)
+    daytona_api_url: str = Field(default='https://app.daytona.io/api')
+    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
 

openhands/core/config/condenser_config.py

@@ -26,8 +26,10 @@ class RecentEventsCondenserConfig(BaseModel):
     """Configuration for RecentEventsCondenser."""
 
     type: Literal['recent'] = Field('recent')
+
+    # at least one event by default, because the best guess is that it is the user task
     keep_first: int = Field(
-        default=0,
+        default=1,
         description='The number of initial events to condense.',
         ge=0,
     )
@@ -43,6 +45,8 @@ class LLMSummarizingCondenserConfig(BaseModel):
     llm_config: LLMConfig = Field(
         ..., description='Configuration for the LLM to use for condensing.'
     )
+
+    # 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.',
@@ -62,8 +66,10 @@ class AmortizedForgettingCondenserConfig(BaseModel):
         description='Maximum size of the condensed history before triggering forgetting.',
         ge=2,
     )
+
+    # at least one event by default, because the best guess is that it's the user task
     keep_first: int = Field(
-        default=0,
+        default=1,
         description='Number of initial events to always keep in history.',
         ge=0,
     )
@@ -81,8 +87,10 @@ class LLMAttentionCondenserConfig(BaseModel):
         description='Maximum size of the condensed history before triggering forgetting.',
         ge=2,
     )
+
+    # at least one event by default, because the best guess is that it's the user task
     keep_first: int = Field(
-        default=0,
+        default=1,
         description='Number of initial events to always keep in history.',
         ge=0,
     )

openhands/core/config/config_utils.py

@@ -25,14 +25,20 @@ def get_field_info(field: FieldInfo) -> dict[str, Any]:
     # Note: this only works for UnionTypes with None as one of the types
     if get_origin(field_type) is UnionType:
         types = get_args(field_type)
-        non_none_arg = next((t for t in types if t is not type(None)), None)
+        non_none_arg = next(
+            (t for t in types if t is not None and t is not type(None)), None
+        )
         if non_none_arg is not None:
             field_type = non_none_arg
             optional = True
 
     # type name in a pretty format
     type_name = (
-        field_type.__name__ if hasattr(field_type, '__name__') else str(field_type)
+        str(field_type)
+        if field_type is None
+        else (
+            field_type.__name__ if hasattr(field_type, '__name__') else str(field_type)
+        )
     )
 
     # default is always present

openhands/core/config/extended_config.py

@@ -0,0 +1,40 @@
+from pydantic import RootModel
+
+
+class ExtendedConfig(RootModel[dict]):
+    """Configuration for extended functionalities.
+
+    This is implemented as a root model so that the entire input is stored
+    as the root value. This allows arbitrary keys to be stored and later
+    accessed via attribute or dictionary-style access.
+    """
+
+    @property
+    def root(self) -> dict:  # type annotation to help mypy
+        return super().root
+
+    def __str__(self) -> str:
+        # Use the root dict to build a string representation.
+        attr_str = [f'{k}={repr(v)}' for k, v in self.root.items()]
+        return f"ExtendedConfig({', '.join(attr_str)})"
+
+    def __repr__(self) -> str:
+        return self.__str__()
+
+    @classmethod
+    def from_dict(cls, data: dict) -> 'ExtendedConfig':
+        # Create an instance directly by wrapping the input dict.
+        return cls(data)
+
+    def __getitem__(self, key: str) -> object:
+        # Provide dictionary-like access via the root dict.
+        return self.root[key]
+
+    def __getattr__(self, key: str) -> object:
+        # Fallback for attribute access using the root dict.
+        try:
+            return self.root[key]
+        except KeyError as e:
+            raise AttributeError(
+                f"'ExtendedConfig' object has no attribute '{key}'"
+            ) from e

openhands/core/config/sandbox_config.py

@@ -53,11 +53,11 @@ class SandboxConfig(BaseModel):
     remote_runtime_api_timeout: int = Field(default=10)
     remote_runtime_enable_retries: bool = Field(default=False)
     remote_runtime_class: str | None = Field(
-        default='sysbox'
+        default=None
     )  # can be "None" (default to gvisor) or "sysbox" (support docker inside runtime + more stable)
     enable_auto_lint: bool = Field(
-        default=False  # once enabled, OpenHands would lint files after editing
-    )
+        default=False
+    )  # once enabled, OpenHands would lint files after editing
     use_host_network: bool = Field(default=False)
     runtime_extra_build_args: list[str] | None = Field(default=None)
     initialize_plugins: bool = Field(default=True)

openhands/core/config/utils.py

@@ -19,6 +19,7 @@ from openhands.core.config.config_utils import (
     OH_DEFAULT_AGENT,
     OH_MAX_ITERATIONS,
 )
+from openhands.core.config.extended_config import ExtendedConfig
 from openhands.core.config.llm_config import LLMConfig
 from openhands.core.config.sandbox_config import SandboxConfig
 from openhands.core.config.security_config import SecurityConfig
@@ -134,6 +135,10 @@ def load_from_toml(cfg: AppConfig, toml_file: str = 'config.toml') -> None:
     for key, value in toml_config.items():
         if isinstance(value, dict):
             try:
+                if key.lower() == 'extended':
+                    # For ExtendedConfig (RootModel), pass the entire dict as the root value
+                    cfg.extended = ExtendedConfig(value)
+                    continue
                 if key is not None and key.lower() == 'agent':
                     # Every entry here is either a field for the default `agent` config group, or itself a group
                     # The best way to tell the difference is to try to parse it as an AgentConfig object

openhands/core/exceptions.py

@@ -10,17 +10,17 @@ class AgentError(Exception):
 
 
 class AgentNoInstructionError(AgentError):
-    def __init__(self, message='Instruction must be provided'):
+    def __init__(self, message: str = 'Instruction must be provided') -> None:
         super().__init__(message)
 
 
 class AgentEventTypeError(AgentError):
-    def __init__(self, message='Event must be a dictionary'):
+    def __init__(self, message: str = 'Event must be a dictionary') -> None:
         super().__init__(message)
 
 
 class AgentAlreadyRegisteredError(AgentError):
-    def __init__(self, name=None):
+    def __init__(self, name: str | None = None) -> None:
         if name is not None:
             message = f"Agent class already registered under '{name}'"
         else:
@@ -29,7 +29,7 @@ class AgentAlreadyRegisteredError(AgentError):
 
 
 class AgentNotRegisteredError(AgentError):
-    def __init__(self, name=None):
+    def __init__(self, name: str | None = None) -> None:
         if name is not None:
             message = f"No agent class registered under '{name}'"
         else:
@@ -38,7 +38,7 @@ class AgentNotRegisteredError(AgentError):
 
 
 class AgentStuckInLoopError(AgentError):
-    def __init__(self, message='Agent got stuck in a loop'):
+    def __init__(self, message: str = 'Agent got stuck in a loop') -> None:
         super().__init__(message)
 
 
@@ -48,7 +48,7 @@ class AgentStuckInLoopError(AgentError):
 
 
 class TaskInvalidStateError(Exception):
-    def __init__(self, state=None):
+    def __init__(self, state: str | None = None) -> None:
         if state is not None:
             message = f'Invalid state {state}'
         else:
@@ -64,45 +64,47 @@ class TaskInvalidStateError(Exception):
 # This exception gets sent back to the LLM
 # It might be malformed JSON
 class LLMMalformedActionError(Exception):
-    def __init__(self, message='Malformed response'):
+    def __init__(self, message: str = 'Malformed response') -> None:
         self.message = message
         super().__init__(message)
 
-    def __str__(self):
+    def __str__(self) -> str:
         return self.message
 
 
 # This exception gets sent back to the LLM
 # For some reason, the agent did not return an action
 class LLMNoActionError(Exception):
-    def __init__(self, message='Agent must return an action'):
+    def __init__(self, message: str = 'Agent must return an action') -> None:
         super().__init__(message)
 
 
 # This exception gets sent back to the LLM
 # The LLM output did not include an action, or the action was not the expected type
 class LLMResponseError(Exception):
-    def __init__(self, message='Failed to retrieve action from LLM response'):
+    def __init__(
+        self, message: str = 'Failed to retrieve action from LLM response'
+    ) -> None:
         super().__init__(message)
 
 
 class UserCancelledError(Exception):
-    def __init__(self, message='User cancelled the request'):
+    def __init__(self, message: str = 'User cancelled the request') -> None:
         super().__init__(message)
 
 
 class OperationCancelled(Exception):
     """Exception raised when an operation is cancelled (e.g. by a keyboard interrupt)."""
 
-    def __init__(self, message='Operation was cancelled'):
+    def __init__(self, message: str = 'Operation was cancelled') -> None:
         super().__init__(message)
 
 
 class LLMContextWindowExceedError(RuntimeError):
     def __init__(
         self,
-        message='Conversation history longer than LLM context window limit. Consider turning on enable_history_truncation config to avoid this error',
-    ):
+        message: str = 'Conversation history longer than LLM context window limit. Consider turning on enable_history_truncation config to avoid this error',
+    ) -> None:
         super().__init__(message)
 
 
@@ -117,7 +119,7 @@ class FunctionCallConversionError(Exception):
     This typically happens when there's a malformed message (e.g., missing <function=...> tags). But not due to LLM output.
     """
 
-    def __init__(self, message):
+    def __init__(self, message: str) -> None:
         super().__init__(message)
 
 
@@ -127,14 +129,14 @@ class FunctionCallValidationError(Exception):
     This typically happens when the LLM outputs unrecognized function call / parameter names / values.
     """
 
-    def __init__(self, message):
+    def __init__(self, message: str) -> None:
         super().__init__(message)
 
 
 class FunctionCallNotExistsError(Exception):
     """Exception raised when an LLM call a tool that is not registered."""
 
-    def __init__(self, message):
+    def __init__(self, message: str) -> None:
         super().__init__(message)
 
 
@@ -191,15 +193,17 @@ class AgentRuntimeNotFoundError(AgentRuntimeUnavailableError):
 
 
 class BrowserInitException(Exception):
-    def __init__(self, message='Failed to initialize browser environment'):
+    def __init__(
+        self, message: str = 'Failed to initialize browser environment'
+    ) -> None:
         super().__init__(message)
 
 
 class BrowserUnavailableException(Exception):
     def __init__(
         self,
-        message='Browser environment is not available, please check if has been initialized',
-    ):
+        message: str = 'Browser environment is not available, please check if has been initialized',
+    ) -> None:
         super().__init__(message)
 
 
@@ -217,5 +221,5 @@ class MicroAgentError(Exception):
 class MicroAgentValidationError(MicroAgentError):
     """Raised when there's a validation error in microagent metadata."""
 
-    def __init__(self, message='Micro agent validation failed'):
+    def __init__(self, message: str = 'Micro agent validation failed') -> None:
         super().__init__(message)

openhands/core/logger.py

@@ -74,10 +74,11 @@ LOG_COLORS: Mapping[str, ColorType] = {
 
 
 class StackInfoFilter(logging.Filter):
-    def filter(self, record):
+    def filter(self, record: logging.LogRecord) -> bool:
         if record.levelno >= logging.ERROR:
-            record.stack_info = True
-            record.exc_info = True
+            # LogRecord attributes are dynamically typed
+            setattr(record, 'stack_info', True)
+            setattr(record, 'exc_info', sys.exc_info())
         return True
 
 
@@ -107,9 +108,9 @@ def strip_ansi(s: str) -> str:
 
 
 class ColoredFormatter(logging.Formatter):
-    def format(self, record):
-        msg_type = record.__dict__.get('msg_type')
-        event_source = record.__dict__.get('event_source')
+    def format(self, record: logging.LogRecord) -> str:
+        msg_type = record.__dict__.get('msg_type', '')
+        event_source = record.__dict__.get('event_source', '')
         if event_source:
             new_msg_type = f'{event_source.upper()}_{msg_type}'
             if new_msg_type in LOG_COLORS:
@@ -136,12 +137,13 @@ class ColoredFormatter(logging.Formatter):
         return super().format(new_record)
 
 
-def _fix_record(record: logging.LogRecord):
+def _fix_record(record: logging.LogRecord) -> logging.LogRecord:
     new_record = copy.copy(record)
     # The formatter expects non boolean values, and will raise an exception if there is a boolean - so we fix these
-    if new_record.exc_info is True and not new_record.exc_text:  # type: ignore
-        new_record.exc_info = sys.exc_info()  # type: ignore
-        new_record.stack_info = None  # type: ignore
+    # LogRecord attributes are dynamically typed
+    if getattr(new_record, 'exc_info', None) is True:
+        setattr(new_record, 'exc_info', sys.exc_info())
+        setattr(new_record, 'stack_info', None)
     return new_record
 
 
@@ -158,32 +160,32 @@ class RollingLogger:
     log_lines: list[str]
     all_lines: str
 
-    def __init__(self, max_lines=10, char_limit=80):
+    def __init__(self, max_lines: int = 10, char_limit: int = 80) -> None:
         self.max_lines = max_lines
         self.char_limit = char_limit
         self.log_lines = [''] * self.max_lines
         self.all_lines = ''
 
-    def is_enabled(self):
+    def is_enabled(self) -> bool:
         return DEBUG and sys.stdout.isatty()
 
-    def start(self, message=''):
+    def start(self, message: str = '') -> None:
         if message:
             print(message)
         self._write('\n' * self.max_lines)
         self._flush()
 
-    def add_line(self, line):
+    def add_line(self, line: str) -> None:
         self.log_lines.pop(0)
         self.log_lines.append(line[: self.char_limit])
         self.print_lines()
         self.all_lines += line + '\n'
 
-    def write_immediately(self, line):
+    def write_immediately(self, line: str) -> None:
         self._write(line)
         self._flush()
 
-    def print_lines(self):
+    def print_lines(self) -> None:
         """Display the last n log_lines in the console (not for file logging).
 
         This will create the effect of a rolling display in the console.
@@ -192,31 +194,31 @@ class RollingLogger:
         for line in self.log_lines:
             self.replace_current_line(line)
 
-    def move_back(self, amount=-1):
+    def move_back(self, amount: int = -1) -> None:
         r"""'\033[F' moves the cursor up one line."""
         if amount == -1:
             amount = self.max_lines
         self._write('\033[F' * (self.max_lines))
         self._flush()
 
-    def replace_current_line(self, line=''):
+    def replace_current_line(self, line: str = '') -> None:
         r"""'\033[2K\r' clears the line and moves the cursor to the beginning of the line."""
         self._write('\033[2K' + line + '\n')
         self._flush()
 
-    def _write(self, line):
+    def _write(self, line: str) -> None:
         if not self.is_enabled():
             return
         sys.stdout.write(line)
 
-    def _flush(self):
+    def _flush(self) -> None:
         if not self.is_enabled():
             return
         sys.stdout.flush()
 
 
 class SensitiveDataFilter(logging.Filter):
-    def filter(self, record):
+    def filter(self, record: logging.LogRecord) -> bool:
         # Gather sensitive values which should not ever appear in the logs.
         sensitive_values = []
         for key, value in os.environ.items():
@@ -245,6 +247,7 @@ class SensitiveDataFilter(logging.Filter):
             'modal_api_token_secret',
             'llm_api_key',
             'sandbox_env_github_token',
+            'daytona_api_key',
         ]
 
         # add env var names
@@ -262,7 +265,9 @@ class SensitiveDataFilter(logging.Filter):
         return True
 
 
-def get_console_handler(log_level: int = logging.INFO, extra_info: str | None = None):
+def get_console_handler(
+    log_level: int = logging.INFO, extra_info: str | None = None
+) -> logging.StreamHandler:
     """Returns a console handler for logging."""
     console_handler = logging.StreamHandler()
     console_handler.setLevel(log_level)
@@ -273,7 +278,9 @@ def get_console_handler(log_level: int = logging.INFO, extra_info: str | None =
     return console_handler
 
 
-def get_file_handler(log_dir: str, log_level: int = logging.INFO):
+def get_file_handler(
+    log_dir: str, log_level: int = logging.INFO
+) -> logging.FileHandler:
     """Returns a file handler for logging."""
     os.makedirs(log_dir, exist_ok=True)
     timestamp = datetime.now().strftime('%Y-%m-%d')
@@ -347,7 +354,13 @@ logging.getLogger('LiteLLM Proxy').disabled = True
 class LlmFileHandler(logging.FileHandler):
     """LLM prompt and response logging."""
 
-    def __init__(self, filename, mode='a', encoding='utf-8', delay=False):
+    def __init__(
+        self,
+        filename: str,
+        mode: str = 'a',
+        encoding: str = 'utf-8',
+        delay: bool = False,
+    ) -> None:
         """Initializes an instance of LlmFileHandler.
 
         Args:
@@ -378,7 +391,7 @@ class LlmFileHandler(logging.FileHandler):
         self.baseFilename = os.path.join(self.log_directory, filename)
         super().__init__(self.baseFilename, mode, encoding, delay)
 
-    def emit(self, record):
+    def emit(self, record: logging.LogRecord) -> None:
         """Emits a log record.
 
         Args:
@@ -393,7 +406,7 @@ class LlmFileHandler(logging.FileHandler):
         self.message_counter += 1
 
 
-def _get_llm_file_handler(name: str, log_level: int):
+def _get_llm_file_handler(name: str, log_level: int) -> LlmFileHandler:
     # The 'delay' parameter, when set to True, postpones the opening of the log file
     # until the first log message is emitted.
     llm_file_handler = LlmFileHandler(name, delay=True)
@@ -402,7 +415,7 @@ def _get_llm_file_handler(name: str, log_level: int):
     return llm_file_handler
 
 
-def _setup_llm_logger(name: str, log_level: int):
+def _setup_llm_logger(name: str, log_level: int) -> logging.Logger:
     logger = logging.getLogger(name)
     logger.propagate = False
     logger.setLevel(log_level)

openhands/core/message.py

@@ -15,7 +15,9 @@ class Content(BaseModel):
     cache_prompt: bool = False
 
     @model_serializer
-    def serialize_model(self):
+    def serialize_model(
+        self,
+    ) -> dict[str, str | dict[str, str]] | list[dict[str, str | dict[str, str]]]:
         raise NotImplementedError('Subclasses should implement this method.')
 
 
@@ -24,7 +26,7 @@ class TextContent(Content):
     text: str
 
     @model_serializer
-    def serialize_model(self):
+    def serialize_model(self) -> dict[str, str | dict[str, str]]:
         data: dict[str, str | dict[str, str]] = {
             'type': self.type,
             'text': self.text,
@@ -39,7 +41,7 @@ class ImageContent(Content):
     image_urls: list[str]
 
     @model_serializer
-    def serialize_model(self):
+    def serialize_model(self) -> list[dict[str, str | dict[str, str]]]:
         images: list[dict[str, str | dict[str, str]]] = []
         for url in self.image_urls:
             images.append({'type': self.type, 'image_url': {'url': url}})
@@ -101,15 +103,22 @@ class Message(BaseModel):
             # See discussion here for details: https://github.com/BerriAI/litellm/issues/6422#issuecomment-2438765472
             if self.role == 'tool' and item.cache_prompt:
                 role_tool_with_prompt_caching = True
-                if isinstance(d, dict):
-                    d.pop('cache_control')
-                elif isinstance(d, list):
-                    for d_item in d:
-                        d_item.pop('cache_control')
+                if isinstance(item, TextContent):
+                    d.pop('cache_control', None)
+                elif isinstance(item, ImageContent):
+                    # ImageContent.model_dump() always returns a list
+                    # We know d is a list of dicts for ImageContent
+                    if hasattr(d, '__iter__'):
+                        for d_item in d:
+                            if hasattr(d_item, 'pop'):
+                                d_item.pop('cache_control', None)
+
             if isinstance(item, TextContent):
                 content.append(d)
             elif isinstance(item, ImageContent) and self.vision_enabled:
-                content.extend(d)
+                # ImageContent.model_dump() always returns a list
+                # We know d is a list for ImageContent
+                content.extend([d] if isinstance(d, dict) else d)
 
         message_dict: dict = {'content': content, 'role': self.role}
 

openhands/core/message_utils.py

@@ -29,6 +29,7 @@ from openhands.events.observation import (
 from openhands.events.observation.error import ErrorObservation
 from openhands.events.observation.observation import Observation
 from openhands.events.serialization.event import truncate_content
+from openhands.llm.metrics import Metrics, TokenUsage
 
 
 def events_to_messages(
@@ -159,7 +160,7 @@ def get_action_message(
         )
 
         llm_response: ModelResponse = tool_metadata.model_response
-        assistant_msg = llm_response.choices[0].message
+        assistant_msg = getattr(llm_response.choices[0], 'message')
 
         # Add the LLM message (assistant) that initiated the tool calls
         # (overwrites any previous message with the same response_id)
@@ -167,7 +168,7 @@ def get_action_message(
             f'Tool calls type: {type(assistant_msg.tool_calls)}, value: {assistant_msg.tool_calls}'
         )
         pending_tool_call_action_messages[llm_response.id] = Message(
-            role=assistant_msg.role,
+            role=getattr(assistant_msg, 'role', 'assistant'),
             # tool call content SHOULD BE a string
             content=[TextContent(text=assistant_msg.content or '')]
             if assistant_msg.content is not None
@@ -184,7 +185,7 @@ def get_action_message(
         tool_metadata = action.tool_call_metadata
         if tool_metadata is not None:
             # take the response message from the tool call
-            assistant_msg = tool_metadata.model_response.choices[0].message
+            assistant_msg = getattr(tool_metadata.model_response.choices[0], 'message')
             content = assistant_msg.content or ''
 
             # save content if any, to thought
@@ -196,9 +197,11 @@ def get_action_message(
 
             # remove the tool call metadata
             action.tool_call_metadata = None
+        if role not in ('user', 'system', 'assistant', 'tool'):
+            raise ValueError(f'Invalid role: {role}')
         return [
             Message(
-                role=role,
+                role=role,  # type: ignore[arg-type]
                 content=[TextContent(text=action.thought)],
             )
         ]
@@ -207,9 +210,11 @@ def get_action_message(
         content = [TextContent(text=action.content or '')]
         if vision_is_active and action.image_urls:
             content.append(ImageContent(image_urls=action.image_urls))
+        if role not in ('user', 'system', 'assistant', 'tool'):
+            raise ValueError(f'Invalid role: {role}')
         return [
             Message(
-                role=role,
+                role=role,  # type: ignore[arg-type]
                 content=content,
             )
         ]
@@ -217,7 +222,7 @@ def get_action_message(
         content = [TextContent(text=f'User executed the command:\n{action.command}')]
         return [
             Message(
-                role='user',
+                role='user',  # Always user for CmdRunAction
                 content=content,
             )
         ]
@@ -362,3 +367,47 @@ def apply_prompt_caching(messages: list[Message]) -> None:
                 -1
             ].cache_prompt = True  # Last item inside the message content
             break
+
+
+def get_token_usage_for_event(event: Event, metrics: Metrics) -> TokenUsage | None:
+    """
+    Returns at most one token usage record for the `model_response.id` in this event's
+    `tool_call_metadata`.
+
+    If no response_id is found, or none match in metrics.token_usages, returns None.
+    """
+    if event.tool_call_metadata and event.tool_call_metadata.model_response:
+        response_id = event.tool_call_metadata.model_response.get('id')
+        if response_id:
+            return next(
+                (
+                    usage
+                    for usage in metrics.token_usages
+                    if usage.response_id == response_id
+                ),
+                None,
+            )
+    return None
+
+
+def get_token_usage_for_event_id(
+    events: list[Event], event_id: int, metrics: Metrics
+) -> TokenUsage | None:
+    """
+    Starting from the event with .id == event_id and moving backwards in `events`,
+    find the first TokenUsage record (if any) associated with a response_id from
+    tool_call_metadata.model_response.id.
+
+    Returns the first match found, or None if none is found.
+    """
+    # find the index of the event with the given id
+    idx = next((i for i, e in enumerate(events) if e.id == event_id), None)
+    if idx is None:
+        return None
+
+    # search backward from idx down to 0
+    for i in range(idx, -1, -1):
+        usage = get_token_usage_for_event(events[i], metrics)
+        if usage is not None:
+            return usage
+    return None

openhands/llm/llm.py

@@ -42,6 +42,7 @@ LLM_RETRY_EXCEPTIONS: tuple[type[Exception], ...] = (RateLimitError,)
 # cache prompt supporting models
 # remove this when we gemini and deepseek are supported
 CACHE_PROMPT_SUPPORTED_MODELS = [
+    'claude-3-7-sonnet-20250219',
     'claude-3-5-sonnet-20241022',
     'claude-3-5-sonnet-20240620',
     'claude-3-5-haiku-20241022',
@@ -51,6 +52,7 @@ CACHE_PROMPT_SUPPORTED_MODELS = [
 
 # function calling supporting models
 FUNCTION_CALLING_SUPPORTED_MODELS = [
+    'claude-3-7-sonnet-20250219',
     'claude-3-5-sonnet',
     'claude-3-5-sonnet-20240620',
     'claude-3-5-sonnet-20241022',
@@ -497,20 +499,21 @@ class LLM(RetryMixin, DebugMixin):
             stats += 'Response Latency: %.3f seconds\n' % latest_latency.latency
 
         usage: Usage | None = response.get('usage')
+        response_id = response.get('id', 'unknown')
 
         if usage:
             # keep track of the input and output tokens
-            input_tokens = usage.get('prompt_tokens')
-            output_tokens = usage.get('completion_tokens')
+            prompt_tokens = usage.get('prompt_tokens', 0)
+            completion_tokens = usage.get('completion_tokens', 0)
 
-            if input_tokens:
-                stats += 'Input tokens: ' + str(input_tokens)
+            if prompt_tokens:
+                stats += 'Input tokens: ' + str(prompt_tokens)
 
-            if output_tokens:
+            if completion_tokens:
                 stats += (
-                    (' | ' if input_tokens else '')
+                    (' | ' if prompt_tokens else '')
                     + 'Output tokens: '
-                    + str(output_tokens)
+                    + str(completion_tokens)
                     + '\n'
                 )
 
@@ -519,7 +522,7 @@ class LLM(RetryMixin, DebugMixin):
                 'prompt_tokens_details'
             )
             cache_hit_tokens = (
-                prompt_tokens_details.cached_tokens if prompt_tokens_details else None
+                prompt_tokens_details.cached_tokens if prompt_tokens_details else 0
             )
             if cache_hit_tokens:
                 stats += 'Input tokens (cache hit): ' + str(cache_hit_tokens) + '\n'
@@ -528,10 +531,20 @@ class LLM(RetryMixin, DebugMixin):
             # but litellm doesn't separate them in the usage stats
             # so we can read it from the provider-specific extra field
             model_extra = usage.get('model_extra', {})
-            cache_write_tokens = model_extra.get('cache_creation_input_tokens')
+            cache_write_tokens = model_extra.get('cache_creation_input_tokens', 0)
             if cache_write_tokens:
                 stats += 'Input tokens (cache write): ' + str(cache_write_tokens) + '\n'
 
+            # Record in metrics
+            # We'll treat cache_hit_tokens as "cache read" and cache_write_tokens as "cache write"
+            self.metrics.add_token_usage(
+                prompt_tokens=prompt_tokens,
+                completion_tokens=completion_tokens,
+                cache_read_tokens=cache_hit_tokens,
+                cache_write_tokens=cache_write_tokens,
+                response_id=response_id,
+            )
+
         # log the stats
         if stats:
             logger.debug(stats)

openhands/llm/metrics.py

@@ -17,11 +17,23 @@ class ResponseLatency(BaseModel):
     response_id: str
 
 
+class TokenUsage(BaseModel):
+    """Metric tracking detailed token usage per completion call."""
+
+    model: str
+    prompt_tokens: int
+    completion_tokens: int
+    cache_read_tokens: int
+    cache_write_tokens: int
+    response_id: str
+
+
 class Metrics:
     """Metrics class can record various metrics during running and evaluation.
-    Currently, we define the following metrics:
-        accumulated_cost: the total cost (USD $) of the current LLM.
-        response_latency: the time taken for each LLM completion call.
+    We track:
+      - accumulated_cost and costs
+      - A list of ResponseLatency
+      - A list of TokenUsage (one per call).
     """
 
     def __init__(self, model_name: str = 'default') -> None:
@@ -29,6 +41,7 @@ class Metrics:
         self._costs: list[Cost] = []
         self._response_latencies: list[ResponseLatency] = []
         self.model_name = model_name
+        self._token_usages: list[TokenUsage] = []
 
     @property
     def accumulated_cost(self) -> float:
@@ -54,6 +67,16 @@ class Metrics:
     def response_latencies(self, value: list[ResponseLatency]) -> None:
         self._response_latencies = value
 
+    @property
+    def token_usages(self) -> list[TokenUsage]:
+        if not hasattr(self, '_token_usages'):
+            self._token_usages = []
+        return self._token_usages
+
+    @token_usages.setter
+    def token_usages(self, value: list[TokenUsage]) -> None:
+        self._token_usages = value
+
     def add_cost(self, value: float) -> None:
         if value < 0:
             raise ValueError('Added cost cannot be negative.')
@@ -67,10 +90,33 @@ class Metrics:
             )
         )
 
+    def add_token_usage(
+        self,
+        prompt_tokens: int,
+        completion_tokens: int,
+        cache_read_tokens: int,
+        cache_write_tokens: int,
+        response_id: str,
+    ) -> None:
+        """Add a single usage record."""
+        self._token_usages.append(
+            TokenUsage(
+                model=self.model_name,
+                prompt_tokens=prompt_tokens,
+                completion_tokens=completion_tokens,
+                cache_read_tokens=cache_read_tokens,
+                cache_write_tokens=cache_write_tokens,
+                response_id=response_id,
+            )
+        )
+
     def merge(self, other: 'Metrics') -> None:
+        """Merge 'other' metrics into this one."""
         self._accumulated_cost += other.accumulated_cost
         self._costs += other._costs
-        self._response_latencies += other._response_latencies
+        # use the property so older picked objects that lack the field won't crash
+        self.token_usages += other.token_usages
+        self.response_latencies += other.response_latencies
 
     def get(self) -> dict:
         """Return the metrics in a dictionary."""
@@ -80,12 +126,14 @@ class Metrics:
             'response_latencies': [
                 latency.model_dump() for latency in self._response_latencies
             ],
+            'token_usages': [usage.model_dump() for usage in self._token_usages],
         }
 
     def reset(self):
         self._accumulated_cost = 0.0
         self._costs = []
         self._response_latencies = []
+        self._token_usages = []
 
     def log(self):
         """Log the metrics."""

openhands/memory/__init__.py

@@ -1,4 +1,4 @@
 from openhands.memory.condenser import Condenser
-from openhands.memory.memory import LongTermMemory
+from openhands.memory.long_term_memory import LongTermMemory
 
 __all__ = ['LongTermMemory', 'Condenser']

openhands/memory/condenser/impl/llm_attention_condenser.py

@@ -18,7 +18,7 @@ class ImportantEventSelection(BaseModel):
 class LLMAttentionCondenser(RollingCondenser):
     """Rolling condenser strategy that uses an LLM to select the most important events when condensing the history."""
 
-    def __init__(self, llm: LLM, max_size: int = 100, keep_first: int = 0):
+    def __init__(self, llm: LLM, max_size: int = 100, keep_first: int = 1):
         if keep_first >= max_size // 2:
             raise ValueError(
                 f'keep_first ({keep_first}) must be less than half of max_size ({max_size})'

openhands/memory/condenser/impl/recent_events_condenser.py

@@ -8,7 +8,7 @@ from openhands.memory.condenser.condenser import Condenser
 class RecentEventsCondenser(Condenser):
     """A condenser that only keeps a certain number of the most recent events."""
 
-    def __init__(self, keep_first: int = 0, max_events: int = 10):
+    def __init__(self, keep_first: int = 1, max_events: int = 10):
         self.keep_first = keep_first
         self.max_events = max_events
 

openhands/runtime/__init__.py

@@ -1,4 +1,5 @@
 from openhands.core.logger import openhands_logger as logger
+from openhands.runtime.impl.daytona.daytona_runtime import DaytonaRuntime
 from openhands.runtime.impl.docker.docker_runtime import (
     DockerRuntime,
 )
@@ -24,6 +25,8 @@ def get_runtime_cls(name: str):
         return RunloopRuntime
     elif name == 'local':
         return LocalRuntime
+    elif name == 'daytona':
+        return DaytonaRuntime
     else:
         raise ValueError(f'Runtime {name} not supported')
 

openhands/runtime/builder/docker.py

@@ -67,7 +67,7 @@ class DockerRuntimeBuilder(RuntimeBuilder):
         """
         self.docker_client = docker.from_env()
         version_info = self.docker_client.version()
-        server_version = version_info.get('Version', '').replace('-', '.')
+        server_version = version_info.get('Version', '').split('+')[0].replace('-', '.')
         if tuple(map(int, server_version.split('.'))) < (18, 9):
             raise AgentRuntimeBuildError(
                 'Docker server version must be >= 18.09 to use BuildKit'

openhands/runtime/impl/daytona/README.md

@@ -0,0 +1,24 @@
+# Daytona Runtime
+
+[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
+
+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:
+
+```bash
+    RUNTIME="daytona"
+    DAYTONA_API_KEY="<your-api-key>"
+```
+Optionally, if you don't want your sandboxes to default to the US region, set:
+
+```bash
+    DAYTONA_TARGET="eu"
+```
+
+## Documentation
+Read more by visiting our [documentation](https://www.daytona.io/docs/) page.

openhands/runtime/impl/daytona/daytona_runtime.py

@@ -0,0 +1,262 @@
+import json
+from typing import Callable
+
+import tenacity
+from daytona_sdk import (
+    CreateWorkspaceParams,
+    Daytona,
+    DaytonaConfig,
+    SessionExecuteRequest,
+    Workspace,
+)
+
+from openhands.core.config.app_config import AppConfig
+from openhands.events.stream import EventStream
+from openhands.runtime.impl.action_execution.action_execution_client import (
+    ActionExecutionClient,
+)
+from openhands.runtime.plugins.requirement import PluginRequirement
+from openhands.runtime.utils.command import get_action_execution_server_startup_command
+from openhands.utils.async_utils import call_sync_from_async
+from openhands.utils.tenacity_stop import stop_if_should_exit
+
+WORKSPACE_PREFIX = 'openhands-sandbox-'
+
+
+class DaytonaRuntime(ActionExecutionClient):
+    """The DaytonaRuntime class is a DockerRuntime that utilizes Daytona workspace as a runtime environment."""
+
+    _sandbox_port: int = 4444
+    _vscode_port: int = 4445
+
+    def __init__(
+        self,
+        config: AppConfig,
+        event_stream: EventStream,
+        sid: str = 'default',
+        plugins: list[PluginRequirement] | None = None,
+        env_vars: dict[str, str] | None = None,
+        status_callback: Callable | None = None,
+        attach_to_existing: bool = False,
+        headless_mode: bool = True,
+    ):
+        assert config.daytona_api_key, 'Daytona API key is required'
+
+        self.config = config
+        self.sid = sid
+        self.workspace_id = WORKSPACE_PREFIX + sid
+        self.workspace: Workspace | None = None
+        self._vscode_url: str | None = None
+
+        daytona_config = DaytonaConfig(
+            api_key=config.daytona_api_key.get_secret_value(),
+            server_url=config.daytona_api_url,
+            target=config.daytona_target,
+        )
+        self.daytona = Daytona(daytona_config)
+
+        # workspace_base cannot be used because we can't bind mount into a workspace.
+        if self.config.workspace_base is not None:
+            self.log(
+                'warning',
+                'Workspace mounting is not supported in the Daytona runtime.',
+            )
+
+        super().__init__(
+            config,
+            event_stream,
+            sid,
+            plugins,
+            env_vars,
+            status_callback,
+            attach_to_existing,
+            headless_mode,
+        )
+
+    def _get_workspace(self) -> Workspace | None:
+        try:
+            workspace = self.daytona.get_current_workspace(self.workspace_id)
+            self.log(
+                'info', f'Attached to existing workspace with id: {self.workspace_id}'
+            )
+        except Exception:
+            self.log(
+                'warning',
+                f'Failed to attach to existing workspace with id: {self.workspace_id}',
+            )
+            workspace = None
+
+        return workspace
+
+    def _get_creation_env_vars(self) -> dict[str, str]:
+        env_vars: dict[str, str] = {
+            'port': str(self._sandbox_port),
+            'PYTHONUNBUFFERED': '1',
+            'VSCODE_PORT': str(self._vscode_port),
+        }
+
+        if self.config.debug:
+            env_vars['DEBUG'] = 'true'
+
+        return env_vars
+
+    def _create_workspace(self) -> Workspace:
+        workspace_params = CreateWorkspaceParams(
+            id=self.workspace_id,
+            language='python',
+            image=self.config.sandbox.runtime_container_image,
+            public=True,
+            env_vars=self._get_creation_env_vars(),
+        )
+        workspace = self.daytona.create(workspace_params)
+        return workspace
+
+    def _get_workspace_status(self) -> str:
+        assert self.workspace is not None, 'Workspace is not initialized'
+        assert (
+            self.workspace.instance.info is not None
+        ), 'Workspace info is not available'
+        assert (
+            self.workspace.instance.info.provider_metadata is not None
+        ), 'Provider metadata is not available'
+
+        provider_metadata = json.loads(self.workspace.instance.info.provider_metadata)
+        return provider_metadata.get('status', 'unknown')
+
+    def _construct_api_url(self, port: int) -> str:
+        assert self.workspace is not None, 'Workspace is not initialized'
+        assert (
+            self.workspace.instance.info is not None
+        ), 'Workspace info is not available'
+        assert (
+            self.workspace.instance.info.provider_metadata is not None
+        ), 'Provider metadata is not available'
+
+        node_domain = json.loads(self.workspace.instance.info.provider_metadata)[
+            'nodeDomain'
+        ]
+        return f'https://{port}-{self.workspace.id}.{node_domain}'
+
+    def _get_action_execution_server_host(self) -> str:
+        return self.api_url
+
+    def _start_action_execution_server(self) -> None:
+        assert self.workspace is not None, 'Workspace is not initialized'
+
+        self.workspace.process.exec(
+            f'mkdir -p {self.config.workspace_mount_path_in_sandbox}'
+        )
+
+        start_command: list[str] = get_action_execution_server_startup_command(
+            server_port=self._sandbox_port,
+            plugins=self.plugins,
+            app_config=self.config,
+            override_user_id=1000,
+            override_username='openhands',
+        )
+        start_command_str: str = ' '.join(start_command)
+
+        self.log(
+            'debug',
+            f'Starting action execution server with command: {start_command_str}',
+        )
+
+        exec_session_id = 'action-execution-server'
+        self.workspace.process.create_session(exec_session_id)
+        self.workspace.process.execute_session_command(
+            exec_session_id,
+            SessionExecuteRequest(command='cd /openhands/code', var_async=True),
+        )
+
+        exec_command = self.workspace.process.execute_session_command(
+            exec_session_id,
+            SessionExecuteRequest(command=start_command_str, var_async=True),
+        )
+
+        self.log('debug', f'exec_command_id: {exec_command.cmd_id}')
+
+    @tenacity.retry(
+        stop=tenacity.stop_after_delay(120) | stop_if_should_exit(),
+        wait=tenacity.wait_fixed(1),
+        reraise=(ConnectionRefusedError,),
+    )
+    def _wait_until_alive(self):
+        super().check_if_alive()
+
+    async def connect(self):
+        self.send_status_message('STATUS$STARTING_RUNTIME')
+
+        if self.attach_to_existing:
+            self.workspace = await call_sync_from_async(self._get_workspace)
+
+        if self.workspace is None:
+            self.send_status_message('STATUS$PREPARING_CONTAINER')
+            self.workspace = await call_sync_from_async(self._create_workspace)
+            self.log('info', f'Created new workspace with id: {self.workspace_id}')
+
+        if self._get_workspace_status() == 'stopped':
+            self.log('info', 'Starting Daytona workspace...')
+            await call_sync_from_async(self.workspace.start)
+
+        self.api_url = await call_sync_from_async(
+            self._construct_api_url, self._sandbox_port
+        )
+
+        if not self.attach_to_existing:
+            await call_sync_from_async(self._start_action_execution_server)
+            self.log(
+                'info',
+                f'Container started. Action execution server url: {self.api_url}',
+            )
+
+        self.log('info', 'Waiting for client to become ready...')
+        self.send_status_message('STATUS$WAITING_FOR_CLIENT')
+        await call_sync_from_async(self._wait_until_alive)
+
+        if not self.attach_to_existing:
+            await call_sync_from_async(self.setup_initial_env)
+
+        self.log(
+            'info',
+            f'Container initialized with plugins: {[plugin.name for plugin in self.plugins]}',
+        )
+
+        if not self.attach_to_existing:
+            self.send_status_message(' ')
+        self._runtime_initialized = True
+
+    def close(self):
+        super().close()
+
+        if self.attach_to_existing:
+            return
+
+        if self.workspace:
+            self.daytona.remove(self.workspace)
+
+    @property
+    def vscode_url(self) -> str | None:
+        if self._vscode_url is not None:  # cached value
+            return self._vscode_url
+        token = super().get_vscode_token()
+        if not token:
+            self.log(
+                'warning', 'Failed to get VSCode token while trying to get VSCode URL'
+            )
+            return None
+        if not self.workspace:
+            self.log(
+                'warning', 'Workspace is not initialized while trying to get VSCode URL'
+            )
+            return None
+        self._vscode_url = (
+            self._construct_api_url(self._vscode_port)
+            + f'/?tkn={token}&folder={self.config.workspace_mount_path_in_sandbox}'
+        )
+
+        self.log(
+            'debug',
+            f'VSCode URL: {self._vscode_url}',
+        )
+
+        return self._vscode_url

openhands/security/invariant/analyzer.py

@@ -307,11 +307,17 @@ class InvariantAnalyzer(SecurityAnalyzer):
         new_elements = parse_element(self.trace, event)
         input = [e.model_dump(exclude_none=True) for e in new_elements]  # type: ignore [call-overload]
         self.trace.extend(new_elements)
-        result, err = self.monitor.check(self.input, input)
+        check_result = self.monitor.check(self.input, input)
         self.input.extend(input)
         risk = ActionSecurityRisk.UNKNOWN
-        if err:
-            logger.warning(f'Error checking policy: {err}')
+        
+        if isinstance(check_result, tuple):
+            result, err = check_result
+            if err:
+                logger.warning(f'Error checking policy: {err}')
+                return risk
+        else:
+            logger.warning(f'Error checking policy: {check_result}')
             return risk
 
         risk = self.get_risk(result)

openhands/security/invariant/client.py

@@ -50,7 +50,7 @@ class InvariantClient:
         return None
 
     class _Policy:
-        def __init__(self, invariant):
+        def __init__(self, invariant: 'InvariantClient') -> None:
             self.server = invariant.server
             self.session_id = invariant.session_id
 
@@ -77,7 +77,7 @@ class InvariantClient:
             except (ConnectionError, Timeout, HTTPError) as err:
                 return None, err
 
-        def from_string(self, rule: str):
+        def from_string(self, rule: str) -> 'InvariantClient._Policy':
             policy_id, err = self._create_policy(rule)
             if err:
                 raise err
@@ -97,7 +97,7 @@ class InvariantClient:
                 return None, err
 
     class _Monitor:
-        def __init__(self, invariant):
+        def __init__(self, invariant: 'InvariantClient') -> None:
             self.server = invariant.server
             self.session_id = invariant.session_id
             self.policy = ''
@@ -114,7 +114,7 @@ class InvariantClient:
             except (ConnectionError, Timeout, HTTPError) as err:
                 return None, err
 
-        def from_string(self, rule: str):
+        def from_string(self, rule: str) -> 'InvariantClient._Monitor':
             monitor_id, err = self._create_monitor(rule)
             if err:
                 raise err

openhands/security/invariant/nodes.py

@@ -1,3 +1,4 @@
+from typing import Any, Iterable, Tuple
 from pydantic import BaseModel, Field
 from pydantic.dataclasses import dataclass
 
@@ -10,7 +11,7 @@ class LLM:
 
 class Event(BaseModel):
     metadata: dict | None = Field(
-        default_factory=dict, description='Metadata associated with the event'
+        default_factory=lambda: dict(), description='Metadata associated with the event'
     )
 
 
@@ -30,7 +31,7 @@ class Message(Event):
     content: str | None
     tool_calls: list[ToolCall] | None = None
 
-    def __rich_repr__(self):
+    def __rich_repr__(self) -> Iterable[Any | tuple[Any] | tuple[str, Any] | tuple[str, Any, Any]]:
         # Print on separate line
         yield 'role', self.role
         yield 'content', self.content

poetry.lock

@@ -1507,6 +1507,47 @@ tests-numpy2 = ["Pillow (>=9.4.0)", "absl-py", "decorator", "elasticsearch (<8.0
 torch = ["torch"]
 vision = ["Pillow (>=9.4.0)"]
 
+[[package]]
+name = "daytona-api-client"
+version = "0.13.0"
+description = "Daytona Workspaces"
+optional = false
+python-versions = "*"
+groups = ["main"]
+files = [
+    {file = "daytona_api_client-0.13.0-py3-none-any.whl", hash = "sha256:c4d0dcb89a328c4d0a97d8f076eaf9a00ccc54a8b9f862f4b3302ae887d03c8f"},
+    {file = "daytona_api_client-0.13.0.tar.gz", hash = "sha256:d62b7cb14361b2706df192d2da7dc2b5d02be6fd4259e9433cf2bfdc5807416d"},
+]
+
+[package.dependencies]
+pydantic = ">=2"
+python-dateutil = ">=2.8.2"
+typing-extensions = ">=4.7.1"
+urllib3 = ">=1.25.3,<3.0.0"
+
+[[package]]
+name = "daytona-sdk"
+version = "0.9.1"
+description = "Python SDK for Daytona"
+optional = false
+python-versions = ">=3.7"
+groups = ["main"]
+files = [
+    {file = "daytona_sdk-0.9.1-py3-none-any.whl", hash = "sha256:cce6c90cd3d578747b3c388e24c811cb0b21ad125d34b32836c50059a577a12a"},
+    {file = "daytona_sdk-0.9.1.tar.gz", hash = "sha256:1e2f219f55130fc72d2f14a57d008b8d3e236d45294e0ca51e249106be5ca5de"},
+]
+
+[package.dependencies]
+daytona_api_client = ">=0.13.0,<1.0.0"
+environs = ">=9.5.0,<10.0.0"
+marshmallow = ">=3.19.0,<4.0.0"
+pydantic = ">=2.4.2,<3.0.0"
+python-dateutil = ">=2.8.2,<3.0.0"
+urllib3 = ">=2.0.7,<3.0.0"
+
+[package.extras]
+dev = ["black (>=22.0.0)", "isort (>=5.10.0)", "pydoc-markdown (>=4.8.2)"]
+
 [[package]]
 name = "debugpy"
 version = "1.8.12"
@@ -1731,6 +1772,28 @@ files = [
     {file = "english-words-2.0.1.tar.gz", hash = "sha256:a4105c57493bb757a3d8973fcf8e1dc05e7ca09c836dff467c3fb445f84bc43d"},
 ]
 
+[[package]]
+name = "environs"
+version = "9.5.0"
+description = "simplified environment variable parsing"
+optional = false
+python-versions = ">=3.6"
+groups = ["main"]
+files = [
+    {file = "environs-9.5.0-py2.py3-none-any.whl", hash = "sha256:1e549569a3de49c05f856f40bce86979e7d5ffbbc4398e7f338574c220189124"},
+    {file = "environs-9.5.0.tar.gz", hash = "sha256:a76307b36fbe856bdca7ee9161e6c466fd7fcffc297109a118c59b54e27e30c9"},
+]
+
+[package.dependencies]
+marshmallow = ">=3.0.0"
+python-dotenv = "*"
+
+[package.extras]
+dev = ["dj-database-url", "dj-email-url", "django-cache-url", "flake8 (==4.0.1)", "flake8-bugbear (==21.9.2)", "mypy (==0.910)", "pre-commit (>=2.4,<3.0)", "pytest", "tox"]
+django = ["dj-database-url", "dj-email-url", "django-cache-url"]
+lint = ["flake8 (==4.0.1)", "flake8-bugbear (==21.9.2)", "mypy (==0.910)", "pre-commit (>=2.4,<3.0)"]
+tests = ["dj-database-url", "dj-email-url", "django-cache-url", "pytest"]
+
 [[package]]
 name = "evaluate"
 version = "0.4.3"
@@ -3132,14 +3195,14 @@ zstd = ["zstandard (>=0.18.0)"]
 
 [[package]]
 name = "huggingface-hub"
-version = "0.28.1"
+version = "0.29.0"
 description = "Client library to download and publish models, datasets and other repos on the huggingface.co hub"
 optional = false
 python-versions = ">=3.8.0"
 groups = ["main", "evaluation", "llama-index"]
 files = [
-    {file = "huggingface_hub-0.28.1-py3-none-any.whl", hash = "sha256:aa6b9a3ffdae939b72c464dbb0d7f99f56e649b55c3d52406f49e0a5a620c0a7"},
-    {file = "huggingface_hub-0.28.1.tar.gz", hash = "sha256:893471090c98e3b6efbdfdacafe4052b20b84d59866fb6f54c33d9af18c303ae"},
+    {file = "huggingface_hub-0.29.0-py3-none-any.whl", hash = "sha256:c02daa0b6bafbdacb1320fdfd1dc7151d0940825c88c4ef89837fdb1f6ea0afe"},
+    {file = "huggingface_hub-0.29.0.tar.gz", hash = "sha256:64034c852be270cac16c5743fe1f659b14515a9de6342d6f42cbb2ede191fc80"},
 ]
 
 [package.dependencies]
@@ -4004,14 +4067,14 @@ files = [
 
 [[package]]
 name = "kubernetes"
-version = "32.0.0"
+version = "32.0.1"
 description = "Kubernetes python client"
 optional = false
 python-versions = ">=3.6"
 groups = ["llama-index"]
 files = [
-    {file = "kubernetes-32.0.0-py2.py3-none-any.whl", hash = "sha256:60fd8c29e8e43d9c553ca4811895a687426717deba9c0a66fb2dcc3f5ef96692"},
-    {file = "kubernetes-32.0.0.tar.gz", hash = "sha256:319fa840345a482001ac5d6062222daeb66ec4d1bcb3087402aed685adf0aecb"},
+    {file = "kubernetes-32.0.1-py2.py3-none-any.whl", hash = "sha256:35282ab8493b938b08ab5526c7ce66588232df00ef5e1dbe88a419107dc10998"},
+    {file = "kubernetes-32.0.1.tar.gz", hash = "sha256:42f43d49abd437ada79a79a16bd48a604d3471a117a8347e87db693f2ba0ba28"},
 ]
 
 [package.dependencies]
@@ -4786,7 +4849,7 @@ version = "3.26.1"
 description = "A lightweight library for converting complex datatypes to and from native Python datatypes."
 optional = false
 python-versions = ">=3.9"
-groups = ["evaluation", "llama-index"]
+groups = ["main", "evaluation", "llama-index"]
 files = [
     {file = "marshmallow-3.26.1-py3-none-any.whl", hash = "sha256:3350409f20a70a7e4e11a27661187b77cdcaeb20abca41c1454fe33636bea09c"},
     {file = "marshmallow-3.26.1.tar.gz", hash = "sha256:e6d8affb6cb61d39d26402096dc0aee12d5a26d490a121f118d2e81dc0719dc6"},
@@ -4933,14 +4996,14 @@ urllib3 = "*"
 
 [[package]]
 name = "mistune"
-version = "3.1.1"
+version = "3.1.2"
 description = "A sane and fast Markdown parser with useful plugins and renderers"
 optional = false
 python-versions = ">=3.8"
 groups = ["runtime"]
 files = [
-    {file = "mistune-3.1.1-py3-none-any.whl", hash = "sha256:02106ac2aa4f66e769debbfa028509a275069dcffce0dfa578edd7b991ee700a"},
-    {file = "mistune-3.1.1.tar.gz", hash = "sha256:e0740d635f515119f7d1feb6f9b192ee60f0cc649f80a8f944f905706a21654c"},
+    {file = "mistune-3.1.2-py3-none-any.whl", hash = "sha256:4b47731332315cdca99e0ded46fc0004001c1299ff773dfb48fbe1fd226de319"},
+    {file = "mistune-3.1.2.tar.gz", hash = "sha256:733bf018ba007e8b5f2d3a9eb624034f6ee26c4ea769a98ec533ee111d504dff"},
 ]
 
 [[package]]
@@ -8391,33 +8454,34 @@ pathspec = ">=0.10.1"
 
 [[package]]
 name = "scikit-image"
-version = "0.25.1"
+version = "0.25.2"
 description = "Image processing in Python"
 optional = false
 python-versions = ">=3.10"
 groups = ["evaluation"]
 files = [
-    {file = "scikit_image-0.25.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:40763a3a089617e6f00f92d46b3475368b9783588a165c2aa854da95b66bb4ff"},
-    {file = "scikit_image-0.25.1-cp310-cp310-macosx_12_0_arm64.whl", hash = "sha256:7c6b69f33e5512ee7fc53361b064430f146583f08dc75317667e81d5f8fcd0c6"},
-    {file = "scikit_image-0.25.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d9187347d115776ff0ddba3e5d2a04638d291b1a62e3c315d17b71eea351cde8"},
-    {file = "scikit_image-0.25.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:cdfca713979ad1873a4b55d94bb1eb4bc713f0c10165b261bf6f7e606f44a00c"},
-    {file = "scikit_image-0.25.1-cp310-cp310-win_amd64.whl", hash = "sha256:167fb146de80bb2a1493d1a760a9ac81644a8a5de254c3dd12a95d1b662d819c"},
-    {file = "scikit_image-0.25.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:c1bde2d5f1dfb23b3c72ef9fcdb2dd5f42fa353e8bd606aea63590eba5e79565"},
-    {file = "scikit_image-0.25.1-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:5112d95cccaa45c434e57efc20c1f721ab439e516e2ed49709ddc2afb7c15c70"},
-    {file = "scikit_image-0.25.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0f5e313b028f5d7a9f3888ad825ddf4fb78913d7762891abb267b99244b4dd31"},
-    {file = "scikit_image-0.25.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:39ad76aeff754048dabaff83db752aa0655dee425f006678d14485471bdb459d"},
-    {file = "scikit_image-0.25.1-cp311-cp311-win_amd64.whl", hash = "sha256:8dc8b06176c1a2316fa8bc539fd7e96155721628ae5cf51bc1a2c62cb9786581"},
-    {file = "scikit_image-0.25.1-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:ebf83699d60134909647395a0bf07db3859646de7192b088e656deda6bc15e95"},
-    {file = "scikit_image-0.25.1-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:408086520eed036340e634ab7e4f648e00238f711bac61ab933efeb11464a238"},
-    {file = "scikit_image-0.25.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5bd709faa87795869ccd21f32490c37989ca5846571495822f4b9430fb42c34c"},
-    {file = "scikit_image-0.25.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a6b15c0265c072a46ff4720784d756d8f8e5d63567639aa8451f6673994d6846"},
-    {file = "scikit_image-0.25.1-cp312-cp312-win_amd64.whl", hash = "sha256:a689a0d091e0bd97d7767309abdeb27c43be210d075abb34e71657add920c22b"},
-    {file = "scikit_image-0.25.1-cp313-cp313-macosx_10_13_x86_64.whl", hash = "sha256:f070f899d6572a125ab106c4b26d1a5fb784dc60ba6dea45c7816f08c3a4fb4d"},
-    {file = "scikit_image-0.25.1-cp313-cp313-macosx_12_0_arm64.whl", hash = "sha256:cc9538d8db7670878aa68ea79c0b1796b6c771085e8d50f5408ee617da3281b6"},
-    {file = "scikit_image-0.25.1-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:caa08d4fa851e1f421fcad8eac24d32f2810971dc61f1d72dc950ca9e9ec39b1"},
-    {file = "scikit_image-0.25.1-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c9923aa898b7921fbcf503d32574d48ed937a7cff45ce8587be4868b39676e18"},
-    {file = "scikit_image-0.25.1-cp313-cp313-win_amd64.whl", hash = "sha256:6c7bba6773ab8c39ee8b1cbb17c7f98965bacdb8cd8da337942be6acc38fc562"},
-    {file = "scikit_image-0.25.1.tar.gz", hash = "sha256:d4ab30540d114d37c35fe5c837f89b94aaba2a7643afae8354aa353319e9bbbb"},
+    {file = "scikit_image-0.25.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:d3278f586793176599df6a4cf48cb6beadae35c31e58dc01a98023af3dc31c78"},
+    {file = "scikit_image-0.25.2-cp310-cp310-macosx_12_0_arm64.whl", hash = "sha256:5c311069899ce757d7dbf1d03e32acb38bb06153236ae77fcd820fd62044c063"},
+    {file = "scikit_image-0.25.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:be455aa7039a6afa54e84f9e38293733a2622b8c2fb3362b822d459cc5605e99"},
+    {file = "scikit_image-0.25.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a4c464b90e978d137330be433df4e76d92ad3c5f46a22f159520ce0fdbea8a09"},
+    {file = "scikit_image-0.25.2-cp310-cp310-win_amd64.whl", hash = "sha256:60516257c5a2d2f74387c502aa2f15a0ef3498fbeaa749f730ab18f0a40fd054"},
+    {file = "scikit_image-0.25.2-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:f4bac9196fb80d37567316581c6060763b0f4893d3aca34a9ede3825bc035b17"},
+    {file = "scikit_image-0.25.2-cp311-cp311-macosx_12_0_arm64.whl", hash = "sha256:d989d64ff92e0c6c0f2018c7495a5b20e2451839299a018e0e5108b2680f71e0"},
+    {file = "scikit_image-0.25.2-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b2cfc96b27afe9a05bc92f8c6235321d3a66499995675b27415e0d0c76625173"},
+    {file = "scikit_image-0.25.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:24cc986e1f4187a12aa319f777b36008764e856e5013666a4a83f8df083c2641"},
+    {file = "scikit_image-0.25.2-cp311-cp311-win_amd64.whl", hash = "sha256:b4f6b61fc2db6340696afe3db6b26e0356911529f5f6aee8c322aa5157490c9b"},
+    {file = "scikit_image-0.25.2-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:8db8dd03663112783221bf01ccfc9512d1cc50ac9b5b0fe8f4023967564719fb"},
+    {file = "scikit_image-0.25.2-cp312-cp312-macosx_12_0_arm64.whl", hash = "sha256:483bd8cc10c3d8a7a37fae36dfa5b21e239bd4ee121d91cad1f81bba10cfb0ed"},
+    {file = "scikit_image-0.25.2-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9d1e80107bcf2bf1291acfc0bf0425dceb8890abe9f38d8e94e23497cbf7ee0d"},
+    {file = "scikit_image-0.25.2-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a17e17eb8562660cc0d31bb55643a4da996a81944b82c54805c91b3fe66f4824"},
+    {file = "scikit_image-0.25.2-cp312-cp312-win_amd64.whl", hash = "sha256:bdd2b8c1de0849964dbc54037f36b4e9420157e67e45a8709a80d727f52c7da2"},
+    {file = "scikit_image-0.25.2-cp313-cp313-macosx_10_13_x86_64.whl", hash = "sha256:7efa888130f6c548ec0439b1a7ed7295bc10105458a421e9bf739b457730b6da"},
+    {file = "scikit_image-0.25.2-cp313-cp313-macosx_12_0_arm64.whl", hash = "sha256:dd8011efe69c3641920614d550f5505f83658fe33581e49bed86feab43a180fc"},
+    {file = "scikit_image-0.25.2-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:28182a9d3e2ce3c2e251383bdda68f8d88d9fff1a3ebe1eb61206595c9773341"},
+    {file = "scikit_image-0.25.2-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b8abd3c805ce6944b941cfed0406d88faeb19bab3ed3d4b50187af55cf24d147"},
+    {file = "scikit_image-0.25.2-cp313-cp313-win_amd64.whl", hash = "sha256:64785a8acefee460ec49a354706db0b09d1f325674107d7fa3eadb663fb56d6f"},
+    {file = "scikit_image-0.25.2-cp313-cp313t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:330d061bd107d12f8d68f1d611ae27b3b813b8cdb0300a71d07b1379178dd4cd"},
+    {file = "scikit_image-0.25.2.tar.gz", hash = "sha256:e5a37e6cd4d0c018a7a55b9d601357e3382826d3888c10d0213fc63bff977dde"},
 ]
 
 [package.dependencies]
@@ -8427,16 +8491,16 @@ networkx = ">=3.0"
 numpy = ">=1.24"
 packaging = ">=21"
 pillow = ">=10.1"
-scipy = ">=1.11.2"
+scipy = ">=1.11.4"
 tifffile = ">=2022.8.12"
 
 [package.extras]
-build = ["Cython (>=3.0.8)", "build (>=1.2.1)", "meson-python (>=0.16)", "ninja (>=1.11.1.1)", "numpy (>=2.0)", "pythran (>=0.16)", "setuptools (>=68)", "spin (==0.13)"]
+build = ["Cython (>=3.0.8)", "build (>=1.2.1)", "meson-python (>=0.16)", "ninja (>=1.11.1.1)", "numpy (>=2.0)", "pythran (>=0.16)", "spin (==0.13)"]
 data = ["pooch (>=1.6.0)"]
 developer = ["ipython", "pre-commit", "tomli"]
-docs = ["PyWavelets (>=1.6)", "dask[array] (>=2022.9.2)", "intersphinx-registry (>=0.2411.14)", "ipykernel", "ipywidgets", "kaleido (==0.2.1)", "matplotlib (>=3.7)", "myst-parser", "numpydoc (>=1.7)", "pandas (>=2.0)", "plotly (>=5.20)", "pooch (>=1.6)", "pydata-sphinx-theme (>=0.16)", "pytest-doctestplus", "scikit-learn (>=1.2)", "seaborn (>=0.11)", "sphinx (>=8.0)", "sphinx-copybutton", "sphinx-gallery[parallel] (>=0.18)", "sphinx_design (>=0.5)", "tifffile (>=2022.8.12)"]
-optional = ["PyWavelets (>=1.6)", "SimpleITK", "astropy (>=5.0)", "cloudpickle (>=0.2.1)", "dask[array] (>=2021.1.0,!=2024.8.0)", "matplotlib (>=3.7)", "pooch (>=1.6.0)", "pyamg (>=5.2)", "scikit-learn (>=1.2)"]
-test = ["asv", "numpydoc (>=1.7)", "pooch (>=1.6.0)", "pytest (>=7.0)", "pytest-cov (>=2.11.0)", "pytest-doctestplus", "pytest-faulthandler", "pytest-localserver"]
+docs = ["PyWavelets (>=1.6)", "dask[array] (>=2023.2.0)", "intersphinx-registry (>=0.2411.14)", "ipykernel", "ipywidgets", "kaleido (==0.2.1)", "matplotlib (>=3.7)", "myst-parser", "numpydoc (>=1.7)", "pandas (>=2.0)", "plotly (>=5.20)", "pooch (>=1.6)", "pydata-sphinx-theme (>=0.16)", "pytest-doctestplus", "scikit-learn (>=1.2)", "seaborn (>=0.11)", "sphinx (>=8.0)", "sphinx-copybutton", "sphinx-gallery[parallel] (>=0.18)", "sphinx_design (>=0.5)", "tifffile (>=2022.8.12)"]
+optional = ["PyWavelets (>=1.6)", "SimpleITK", "astropy (>=5.0)", "cloudpickle (>=1.1.1)", "dask[array] (>=2023.2.0)", "matplotlib (>=3.7)", "pooch (>=1.6.0)", "pyamg (>=5.2)", "scikit-learn (>=1.2)"]
+test = ["asv", "numpydoc (>=1.7)", "pooch (>=1.6.0)", "pytest (>=8)", "pytest-cov (>=2.11.0)", "pytest-doctestplus", "pytest-faulthandler", "pytest-localserver"]
 
 [[package]]
 name = "scikit-learn"
@@ -9206,14 +9270,14 @@ files = [
 
 [[package]]
 name = "tifffile"
-version = "2025.1.10"
+version = "2025.2.18"
 description = "Read and write TIFF files"
 optional = false
 python-versions = ">=3.10"
 groups = ["evaluation"]
 files = [
-    {file = "tifffile-2025.1.10-py3-none-any.whl", hash = "sha256:ed24cf4c99fb13b4f5fb29f8a0d5605e60558c950bccbdca2a6470732a27cfb3"},
-    {file = "tifffile-2025.1.10.tar.gz", hash = "sha256:baaf0a3b87bf7ec375fa1537503353f70497eabe1bdde590f2e41cc0346e612f"},
+    {file = "tifffile-2025.2.18-py3-none-any.whl", hash = "sha256:54b36c4d5e5b8d8920134413edfe5a7cfb1c7617bb50cddf7e2772edb7149043"},
+    {file = "tifffile-2025.2.18.tar.gz", hash = "sha256:8d731789e691b468746c1615d989bc550ac93cf753e9210865222e90a5a95d11"},
 ]
 
 [package.dependencies]
@@ -10789,4 +10853,4 @@ testing = ["coverage[toml]", "zope.event", "zope.testing"]
 [metadata]
 lock-version = "2.1"
 python-versions = "^3.12"
-content-hash = "14998d54438fedacad9d82422003f46d0d7721bd50c2f8096657c15dce0f3edd"
+content-hash = "39e0f069346a4d1e52193899989b79ea3e02f81d67fbb2ac0fdc87e70bd1008f"

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "openhands-ai"
-version = "0.25.0"
+version = "0.26.0"
 description = "OpenHands: Code Less, Make More"
 authors = ["OpenHands"]
 license = "MIT"
@@ -76,6 +76,7 @@ stripe = "^11.5.0"
 ipywidgets = "^8.1.5"
 qtconsole = "^5.6.1"
 memory-profiler = "^0.61.0"
+daytona-sdk = "0.9.1"
 
 [tool.poetry.group.llama-index.dependencies]
 llama-index = "*"
@@ -108,7 +109,6 @@ reportlab = "*"
 [tool.coverage.run]
 concurrency = ["gevent"]
 
-
 [tool.poetry.group.runtime.dependencies]
 jupyterlab = "*"
 notebook = "*"
@@ -137,7 +137,6 @@ ignore = ["D1"]
 [tool.ruff.lint.pydocstyle]
 convention = "google"
 
-
 [tool.poetry.group.evaluation.dependencies]
 streamlit = "*"
 whatthepatch = "*"

tests/regression/cases/hello_world/case.yaml

@@ -1,2 +0,0 @@
-timeout: 120  # 2 minutes
-required: true

tests/regression/cases/hello_world/prompt.txt

@@ -1 +0,0 @@
-Create a bash script called hello.sh that prints "hello world"

tests/regression/cases/hello_world/test.sh

@@ -1,28 +0,0 @@
-#!/bin/bash
-set -e
-set +x
-
-echo "checking hello world"
-pwd
-ls -lah
-
-# Check if hello.sh exists
-if [ ! -f hello.sh ]; then
-    echo "hello.sh does not exist"
-    exit 1
-fi
-
-# Check if it's executable
-if [ ! -x hello.sh ]; then
-    echo "hello.sh is not executable"
-    exit 1
-fi
-
-# Run and check output
-output=$(./hello.sh)
-if [ "$output" != "hello world" ]; then
-    echo "Expected 'hello world' but got: $output"
-    exit 1
-fi
-
-exit 0

tests/regression/run.py

@@ -1,158 +0,0 @@
-#!/usr/bin/env python3
-import asyncio
-import os
-import shutil
-import sys
-from pathlib import Path
-from typing import Any, Dict, Optional
-
-import yaml
-
-import openhands.agenthub  # noqa: F401 - import to register agents
-from openhands.core.config import AppConfig
-from openhands.core.main import auto_continue_response, run_controller
-from openhands.events.action import MessageAction
-
-
-def run_test_case(case_dir: Path) -> bool:
-    """Run a single test case.
-
-    Args:
-        case_dir: Path to the test case directory
-
-    Returns:
-        bool: True if test passed, False if failed
-    """
-    case_name = case_dir.name
-    print(f'Running test case: {case_name}')
-
-    # Read case configuration
-    timeout = 120  # Default timeout 2 minutes
-    required = True
-    case_yaml = case_dir / 'case.yaml'
-
-    if case_yaml.exists():
-        with open(case_yaml) as f:
-            config: Optional[Dict[str, Any]] = yaml.safe_load(f)
-            if config:
-                timeout = config.get('timeout', timeout)
-                required = config.get('required', required)
-
-    # Create workspace directory
-    workspace_dir = case_dir / 'workspace'
-    if workspace_dir.exists():
-        # Clean up any existing workspace
-        shutil.rmtree(workspace_dir)
-    workspace_dir.mkdir(exist_ok=True)
-    temp_path = workspace_dir
-    temp_dir = str(workspace_dir)
-    temp_dir_ctx = None
-
-    if not os.getenv('NO_CLEANUP'):
-
-        class WorkspaceCleanup:
-            def __init__(self, workspace_path: Path):
-                self.workspace_path = workspace_path
-
-            def cleanup(self):
-                if self.workspace_path.exists():
-                    shutil.rmtree(self.workspace_path)
-
-        temp_dir_ctx = WorkspaceCleanup(workspace_dir)
-
-    try:
-        # Check if git repo and commit-ish are specified
-        if case_yaml.exists():
-            with open(case_yaml) as f:
-                config = yaml.safe_load(f)
-                if config and 'git' in config:
-                    repo = config['git']
-                    commit = config.get('commit-ish', 'main')
-                    os.system(f'git clone {repo} {temp_dir}')
-                    os.system(f'cd {temp_dir} && git checkout {commit}')
-
-        # Copy prompt and test script
-        shutil.copy2(case_dir / 'prompt.txt', temp_path / 'prompt.txt')
-        shutil.copy2(case_dir / 'test.sh', temp_path / 'test.sh')
-        os.chmod(temp_path / 'test.sh', 0o755)  # Make test.sh executable
-
-        # Read the prompt
-        with open(case_dir / 'prompt.txt') as f:
-            task_str = f.read()
-
-        # Set up OpenHands configuration
-        config = AppConfig()
-        config.name = case_name
-        config.agent_cls = 'CodeActAgent'
-        config.max_budget_per_task = 100
-        config.max_iterations = 100
-        config.cli_multiline_input = False
-        config.config_file = str(Path(__file__).parent.parent.parent / 'config.toml')
-        config.workspace_base = str(temp_path)
-        config.workspace_mount_path = str(temp_path)
-        config.workspace_mount_path_in_sandbox = '/workspace'
-        config.sandbox.keep_runtime_alive = False
-        config.save_trajectory_path = str(temp_path / 'trajectory.json')
-        initial_user_action = MessageAction(content=task_str)
-
-        # Change to temp directory for test execution
-        original_cwd = os.getcwd()
-        os.chdir(temp_dir)
-
-        try:
-            # Run OpenHands
-            asyncio.run(
-                run_controller(
-                    config=config,
-                    initial_user_action=initial_user_action,
-                    fake_user_response_fn=auto_continue_response,
-                    headless_mode=True,
-                )
-            )
-
-            # Run the test script
-            test_result = os.system('./test.sh')
-            if test_result != 0:
-                print(f'Test case {case_name} failed')
-                if required:
-                    return False
-            else:
-                print(f'Test case {case_name} passed')
-                return True
-
-        except Exception as e:
-            print(f'Error running test case {case_name}: {e}')
-            if required:
-                return False
-            return True
-        finally:
-            os.chdir(original_cwd)
-    finally:
-        if temp_dir_ctx is not None:
-            temp_dir_ctx.cleanup()
-
-    return True
-
-
-def main() -> None:
-    """Run all regression tests."""
-    # Find and run all test cases
-    regression_dir = Path(__file__).parent
-    cases_dir = regression_dir / 'cases'
-
-    all_passed = True
-    for case_dir in cases_dir.iterdir():
-        if case_dir.is_dir():
-            if not run_test_case(case_dir):
-                all_passed = False
-
-    if all_passed:
-        print('All tests completed successfully')
-        sys.exit(0)
-    else:
-        print('Some tests failed')
-        sys.exit(1)
-
-
-if __name__ == '__main__':
-    main()

tests/runtime/conftest.py

@@ -11,6 +11,7 @@ from openhands.core.config import AppConfig, load_app_config
 from openhands.core.logger import openhands_logger as logger
 from openhands.events import EventStream
 from openhands.runtime.base import Runtime
+from openhands.runtime.impl.daytona.daytona_runtime import DaytonaRuntime
 from openhands.runtime.impl.docker.docker_runtime import DockerRuntime
 from openhands.runtime.impl.local.local_runtime import LocalRuntime
 from openhands.runtime.impl.remote.remote_runtime import RemoteRuntime
@@ -130,6 +131,8 @@ def get_runtime_classes() -> list[type[Runtime]]:
         return [RemoteRuntime]
     elif runtime.lower() == 'runloop':
         return [RunloopRuntime]
+    elif runtime.lower() == 'daytona':
+        return [DaytonaRuntime]
     else:
         raise ValueError(f'Invalid runtime: {runtime}')
 

tests/unit/test_cli_sid.py

@@ -0,0 +1,101 @@
+import asyncio
+from argparse import Namespace
+from pathlib import Path
+from unittest.mock import AsyncMock, patch
+
+import pytest
+
+from openhands.core.cli import main
+from openhands.core.config import AppConfig
+from openhands.core.schema import AgentState
+from openhands.events.event import EventSource
+from openhands.events.observation import AgentStateChangedObservation
+
+
+@pytest.fixture
+def mock_runtime():
+    with patch('openhands.core.cli.create_runtime') as mock_create_runtime:
+        mock_runtime_instance = AsyncMock()
+        # Mock the event stream with proper async methods
+        mock_runtime_instance.event_stream = AsyncMock()
+        mock_runtime_instance.event_stream.subscribe = AsyncMock()
+        mock_runtime_instance.event_stream.add_event = AsyncMock()
+        # Mock connect method to return immediately
+        mock_runtime_instance.connect = AsyncMock()
+        # Ensure status_callback is None
+        mock_runtime_instance.status_callback = None
+        mock_create_runtime.return_value = mock_runtime_instance
+        yield mock_runtime_instance
+
+
+@pytest.fixture
+def mock_agent():
+    with patch('openhands.core.cli.create_agent') as mock_create_agent:
+        mock_agent_instance = AsyncMock()
+        mock_create_agent.return_value = mock_agent_instance
+        yield mock_agent_instance
+
+
+@pytest.fixture
+def mock_controller():
+    with patch('openhands.core.cli.create_controller') as mock_create_controller:
+        mock_controller_instance = AsyncMock()
+        # Mock run_until_done to finish immediately
+        mock_controller_instance.run_until_done = AsyncMock(return_value=None)
+        mock_create_controller.return_value = (mock_controller_instance, None)
+        yield mock_controller_instance
+
+
+@pytest.fixture
+def task_file(tmp_path: Path) -> Path:
+    # Create a temporary file with our task
+    task_file = tmp_path / 'task.txt'
+    task_file.write_text('Ask me what your task is')
+    return task_file
+
+
+@pytest.fixture
+def mock_config(task_file: Path):
+    with patch('openhands.core.cli.parse_arguments') as mock_parse_args:
+        # Create a proper Namespace with our temporary task file
+        args = Namespace(file=str(task_file), task=None, directory=None)
+        mock_parse_args.return_value = args
+        with patch('openhands.core.cli.setup_config_from_args') as mock_setup_config:
+            mock_config = AppConfig()
+            mock_setup_config.return_value = mock_config
+            yield mock_config
+
+
+@pytest.mark.asyncio
+async def test_cli_session_id_output(
+    mock_runtime, mock_agent, mock_controller, mock_config, capsys
+):
+    # status_callback is set when initializing the runtime
+    mock_controller.status_callback = None
+
+    # Use input patch just for the exit command
+    with patch('builtins.input', return_value='exit'):
+        # Create a task for main
+        main_task = asyncio.create_task(main(asyncio.get_event_loop()))
+
+        # Give it a moment to display the session ID
+        await asyncio.sleep(0.1)
+
+        # Trigger agent state change to STOPPED to end the main loop
+        event = AgentStateChangedObservation(
+            content='Stop', agent_state=AgentState.STOPPED
+        )
+        event._source = EventSource.AGENT
+        await mock_runtime.event_stream.add_event(event)
+
+        # Wait for main to finish with a timeout
+        try:
+            await asyncio.wait_for(main_task, timeout=1.0)
+        except asyncio.TimeoutError:
+            main_task.cancel()
+
+        # Check the output
+        captured = capsys.readouterr()
+        assert 'Session ID:' in captured.out
+        # Also verify that our task message was processed
+        assert 'Ask me what your task is' in str(mock_runtime.mock_calls)

tests/unit/test_condenser.py

@@ -38,7 +38,7 @@ def create_test_event(
     event = Event()
     event._message = message
     event.timestamp = timestamp if timestamp else datetime.now()
-    if id:
+    if id is not None:
         event._id = id
     event._source = EventSource.USER
     return event
@@ -186,13 +186,14 @@ def test_recent_events_condenser():
     assert result == events
 
     # If the max_events are smaller than the number of events, only keep the last few.
-    max_events = 2
+    max_events = 3
     condenser = RecentEventsCondenser(max_events=max_events)
     result = condenser.condensed_history(mock_state)
 
     assert len(result) == max_events
-    assert result[0]._message == 'Event 4'
-    assert result[1]._message == 'Event 5'
+    assert result[0]._message == 'Event 1'  # kept from keep_first
+    assert result[1]._message == 'Event 4'  # kept from max_events
+    assert result[2]._message == 'Event 5'  # kept from max_events
 
     # If the keep_first flag is set, the first event will always be present.
     keep_first = 1
@@ -211,9 +212,9 @@ def test_recent_events_condenser():
     result = condenser.condensed_history(mock_state)
 
     assert len(result) == max_events
-    assert result[0]._message == 'Event 1'
-    assert result[1]._message == 'Event 2'
-    assert result[2]._message == 'Event 5'
+    assert result[0]._message == 'Event 1'  # kept from keep_first
+    assert result[1]._message == 'Event 2'  # kept from keep_first
+    assert result[2]._message == 'Event 5'  # kept from max_events
 
 
 def test_llm_summarization_condenser_from_config():
@@ -539,7 +540,7 @@ def test_llm_attention_condenser_forgets_when_larger_than_max_size(
 ):
     """Test that the LLMAttentionCondenser forgets events when the context grows too large."""
     max_size = 2
-    condenser = LLMAttentionCondenser(max_size=max_size, llm=mock_llm)
+    condenser = LLMAttentionCondenser(max_size=max_size, keep_first=0, llm=mock_llm)
 
     for i in range(max_size * 10):
         event = create_test_event(f'Event {i}', id=i)
@@ -560,7 +561,7 @@ def test_llm_attention_condenser_forgets_when_larger_than_max_size(
 def test_llm_attention_condenser_handles_events_outside_history(mock_llm, mock_state):
     """Test that the LLMAttentionCondenser handles event IDs that aren't from the event history."""
     max_size = 2
-    condenser = LLMAttentionCondenser(max_size=max_size, llm=mock_llm)
+    condenser = LLMAttentionCondenser(max_size=max_size, keep_first=0, llm=mock_llm)
 
     for i in range(max_size * 10):
         event = create_test_event(f'Event {i}', id=i)
@@ -580,7 +581,7 @@ def test_llm_attention_condenser_handles_events_outside_history(mock_llm, mock_s
 def test_llm_attention_condenser_handles_too_many_events(mock_llm, mock_state):
     """Test that the LLMAttentionCondenser handles when the response contains too many event IDs."""
     max_size = 2
-    condenser = LLMAttentionCondenser(max_size=max_size, llm=mock_llm)
+    condenser = LLMAttentionCondenser(max_size=max_size, keep_first=0, llm=mock_llm)
 
     for i in range(max_size * 10):
         event = create_test_event(f'Event {i}', id=i)
@@ -600,7 +601,9 @@ def test_llm_attention_condenser_handles_too_many_events(mock_llm, mock_state):
 def test_llm_attention_condenser_handles_too_few_events(mock_llm, mock_state):
     """Test that the LLMAttentionCondenser handles when the response contains too few event IDs."""
     max_size = 2
-    condenser = LLMAttentionCondenser(max_size=max_size, llm=mock_llm)
+    # Developer note: We must specify keep_first=0 because
+    # keep_first (1) >= max_size//2 (1) is invalid.
+    condenser = LLMAttentionCondenser(max_size=max_size, keep_first=0, llm=mock_llm)
 
     for i in range(max_size * 10):
         event = create_test_event(f'Event {i}', id=i)
@@ -614,3 +617,33 @@ def test_llm_attention_condenser_handles_too_few_events(mock_llm, mock_state):
 
         # The number of results should bounce back and forth between 1, 2, 1, 2, ...
         assert len(results) == (i % 2) + 1
+
+    # Add a new test verifying that keep_first=1 works with max_size > 2
+
+
+def test_llm_attention_condenser_handles_keep_first_for_larger_max_size(
+    mock_llm, mock_state
+):
+    """Test that LLMAttentionCondenser works when keep_first=1 is allowed (must be less than half of max_size)."""
+    max_size = 4  # so keep_first=1 < (max_size // 2) = 2
+    condenser = LLMAttentionCondenser(max_size=max_size, keep_first=1, llm=mock_llm)
+
+    for i in range(max_size * 2):
+        # We append new events, then ensure some are pruned.
+        event = create_test_event(f'Event {i}', id=i)
+        mock_state.history.append(event)
+
+        mock_llm.set_mock_response_content(
+            ImportantEventSelection(ids=[]).model_dump_json()
+        )
+
+        results = condenser.condensed_history(mock_state)
+
+        # We expect that the first event is always kept, and the tail grows until max_size
+        if len(mock_state.history) <= max_size:
+            # No condensation needed yet
+            assert len(results) == len(mock_state.history)
+        else:
+            # The first event is kept, plus some from the tail
+            assert results[0].id == 0
+            assert len(results) <= max_size

tests/unit/test_config.py

@@ -686,6 +686,7 @@ def test_api_keys_repr_str():
         modal_api_token_id='my_modal_api_token_id',
         modal_api_token_secret='my_modal_api_token_secret',
         runloop_api_key='my_runloop_api_key',
+        daytona_api_key='my_daytona_api_key',
     )
     assert 'my_e2b_api_key' not in repr(app_config)
     assert 'my_e2b_api_key' not in str(app_config)
@@ -697,6 +698,8 @@ def test_api_keys_repr_str():
     assert 'my_modal_api_token_secret' not in str(app_config)
     assert 'my_runloop_api_key' not in repr(app_config)
     assert 'my_runloop_api_key' not in str(app_config)
+    assert 'my_daytona_api_key' not in repr(app_config)
+    assert 'my_daytona_api_key' not in str(app_config)
 
     # Check that no other attrs in AppConfig have 'key' or 'token' in their name
     # This will fail when new attrs are added, and attract attention
@@ -705,6 +708,7 @@ def test_api_keys_repr_str():
         'modal_api_token_id',
         'modal_api_token_secret',
         'runloop_api_key',
+        'daytona_api_key',
     ]
     for attr_name in AppConfig.model_fields.keys():
         if (

tests/unit/test_config_extended.py

@@ -0,0 +1,169 @@
+import os
+
+import pytest
+
+from openhands.core.config.app_config import AppConfig
+from openhands.core.config.extended_config import ExtendedConfig
+from openhands.core.config.utils import load_from_toml
+
+
+def test_extended_config_from_dict():
+    """
+    Test that ExtendedConfig.from_dict successfully creates an instance
+    from a dictionary containing arbitrary extra keys.
+    """
+    data = {'foo': 'bar', 'baz': 123, 'flag': True}
+    ext_cfg = ExtendedConfig.from_dict(data)
+
+    # Check that the keys are accessible both as attributes and via __getitem__
+    assert ext_cfg.foo == 'bar'
+    assert ext_cfg['baz'] == 123
+    assert ext_cfg.flag is True
+    # Verify the root dictionary contains all keys
+    assert ext_cfg.root == data
+
+
+def test_extended_config_empty():
+    """
+    Test that an empty ExtendedConfig can be created and accessed.
+    """
+    ext_cfg = ExtendedConfig.from_dict({})
+    assert ext_cfg.root == {}
+
+    # Creating directly should also work
+    ext_cfg2 = ExtendedConfig({})
+    assert ext_cfg2.root == {}
+
+
+def test_extended_config_str_and_repr():
+    """
+    Test that __str__ and __repr__ return the correct string representations
+    of the ExtendedConfig instance.
+    """
+    data = {'alpha': 'test', 'beta': 42}
+    ext_cfg = ExtendedConfig.from_dict(data)
+    string_repr = str(ext_cfg)
+    repr_str = repr(ext_cfg)
+
+    # Ensure the representations include our key/value pairs
+    assert "alpha='test'" in string_repr
+    assert 'beta=42' in string_repr
+
+    # __repr__ should match __str__
+    assert string_repr == repr_str
+
+
+def test_extended_config_getitem_and_getattr():
+    """
+    Test that __getitem__ and __getattr__ can be used to access values
+    in the ExtendedConfig instance.
+    """
+    data = {'key1': 'value1', 'key2': 2}
+    ext_cfg = ExtendedConfig.from_dict(data)
+
+    # Attribute access
+    assert ext_cfg.key1 == 'value1'
+    # Dictionary-style access
+    assert ext_cfg['key2'] == 2
+
+
+def test_extended_config_invalid_key():
+    """
+    Test that accessing a non-existent key via attribute access raises AttributeError.
+    """
+    data = {'existing': 'yes'}
+    ext_cfg = ExtendedConfig.from_dict(data)
+
+    with pytest.raises(AttributeError):
+        _ = ext_cfg.nonexistent
+
+    with pytest.raises(KeyError):
+        _ = ext_cfg['nonexistent']
+
+
+def test_app_config_extended_from_toml(tmp_path: os.PathLike) -> None:
+    """
+    Test that the [extended] section in a TOML file is correctly loaded into
+    AppConfig.extended and that it accepts arbitrary keys.
+    """
+    # Create a temporary TOML file with multiple sections including [extended]
+    config_content = """
+[core]
+workspace_base = "/tmp/workspace"
+
+[llm]
+model = "test-model"
+api_key = "toml-api-key"
+
+[extended]
+custom1 = "custom_value"
+custom2 = 42
+llm = "overridden"  # even a key like 'llm' is accepted in extended
+
+[agent]
+memory_enabled = true
+"""
+    config_file = tmp_path / 'config.toml'
+    config_file.write_text(config_content)
+
+    # Load the TOML into the AppConfig instance
+    config = AppConfig()
+    load_from_toml(config, str(config_file))
+
+    # Verify that extended section is applied
+    assert config.extended.custom1 == 'custom_value'
+    assert config.extended.custom2 == 42
+    # Even though 'llm' is defined in extended, it should not affect the main llm config.
+    assert config.get_llm_config().model == 'test-model'
+
+
+def test_app_config_extended_default(tmp_path: os.PathLike) -> None:
+    """
+    Test that if there is no [extended] section in the TOML file,
+    AppConfig.extended remains its default (empty) ExtendedConfig.
+    """
+    config_content = """
+[core]
+workspace_base = "/tmp/workspace"
+
+[llm]
+model = "test-model"
+api_key = "toml-api-key"
+
+[agent]
+memory_enabled = true
+"""
+    config_file = tmp_path / 'config.toml'
+    config_file.write_text(config_content)
+
+    config = AppConfig()
+    load_from_toml(config, str(config_file))
+
+    # Extended config should be empty
+    assert config.extended.root == {}
+
+
+def test_app_config_extended_random_keys(tmp_path: os.PathLike) -> None:
+    """
+    Test that the extended section accepts arbitrary keys,
+    including ones not defined in any schema.
+    """
+    config_content = """
+[core]
+workspace_base = "/tmp/workspace"
+
+[extended]
+random_key = "random_value"
+another_key = 3.14
+"""
+    config_file = tmp_path / 'config.toml'
+    config_file.write_text(config_content)
+
+    config = AppConfig()
+    load_from_toml(config, str(config_file))
+
+    # Verify that extended config holds the arbitrary keys with correct values.
+    assert config.extended.random_key == 'random_value'
+    assert config.extended.another_key == 3.14
+    # Verify the root dictionary contains all keys
+    assert config.extended.root == {'random_key': 'random_value', 'another_key': 3.14}