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/dummy-agent-test.yml

@@ -19,25 +19,15 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - name: Free Disk Space (Ubuntu)
-        uses: jlumbroso/free-disk-space@main
-        with:
-          # this might remove tools that are actually needed,
-          # if set to "true" but frees about 6 GB
-          tool-cache: true
-          # all of these default to true, but feel free to set to
-          # "false" if necessary for your workflow
-          android: true
-          dotnet: true
-          haskell: true
-          large-packages: true
-          docker-images: false
-          swap-storage: true
       - name: Set up Docker Buildx
         id: buildx
         uses: docker/setup-buildx-action@v3
       - name: Install tmux
         run: sudo apt-get update && sudo apt-get install -y tmux
+      - name: Setup Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: '22.x'
       - name: Install poetry via pipx
         run: pipx install poetry
       - name: Set up Python

.github/workflows/ghcr-build.yml

@@ -41,22 +41,8 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@v4
-      - name: Free Disk Space (Ubuntu)
-        uses: jlumbroso/free-disk-space@main
-        with:
-          # this might remove tools that are actually needed,
-          # if set to "true" but frees about 6 GB
-          tool-cache: true
-          # all of these default to true, but feel free to set to
-          # "false" if necessary for your workflow
-          android: true
-          dotnet: true
-          haskell: true
-          large-packages: true
-          docker-images: false
-          swap-storage: true
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v3.3.0
+        uses: docker/setup-qemu-action@v3.4.0
         with:
           image: tonistiigi/binfmt:latest
       - name: Login to GHCR
@@ -104,22 +90,8 @@ jobs:
     steps:
       - name: Checkout
         uses: actions/checkout@v4
-      - name: Free Disk Space (Ubuntu)
-        uses: jlumbroso/free-disk-space@main
-        with:
-          # this might remove tools that are actually needed,
-          # if set to "true" but frees about 6 GB
-          tool-cache: true
-          # all of these default to true, but feel free to set to
-          # "false" if necessary for your workflow
-          android: true
-          dotnet: true
-          haskell: true
-          large-packages: true
-          docker-images: false
-          swap-storage: true
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v3.3.0
+        uses: docker/setup-qemu-action@v3.4.0
         with:
           image: tonistiigi/binfmt:latest
       - name: Login to GHCR
@@ -219,7 +191,7 @@ jobs:
             exit 1
           fi
 
-  # Run unit tests with the EventStream runtime Docker images as root
+  # Run unit tests with the Docker runtime Docker images as root
   test_runtime_root:
     name: RT Unit Tests (Root)
     needs: [ghcr_build_runtime]
@@ -230,20 +202,6 @@ jobs:
         base_image: ['nikolaik']
     steps:
       - uses: actions/checkout@v4
-      - name: Free Disk Space (Ubuntu)
-        uses: jlumbroso/free-disk-space@main
-        with:
-          # this might remove tools that are actually needed,
-          # if set to "true" but frees about 6 GB
-          tool-cache: true
-          # all of these default to true, but feel free to set to
-          # "false" if necessary for your workflow
-          android: true
-          dotnet: true
-          haskell: true
-          large-packages: true
-          docker-images: false
-          swap-storage: true
       - name: Set up Docker Buildx
         id: buildx
         uses: docker/setup-buildx-action@v3
@@ -275,7 +233,7 @@ jobs:
         run: pipx install poetry
       - name: Install Python dependencies using Poetry
         run: make install-python-dependencies
-      - name: Run runtime tests
+      - name: Run docker runtime tests
         run: |
           # We install pytest-xdist in order to run tests across CPUs
           poetry run pip install pytest-xdist
@@ -286,7 +244,7 @@ jobs:
           image_name=ghcr.io/${{ github.repository_owner }}/runtime:${{ env.RELEVANT_SHA }}-${{ matrix.base_image }}
           image_name=$(echo $image_name | tr '[:upper:]' '[:lower:]')
 
-          TEST_RUNTIME=eventstream \
+          TEST_RUNTIME=docker \
           SANDBOX_USER_ID=$(id -u) \
           SANDBOX_RUNTIME_CONTAINER_IMAGE=$image_name \
           TEST_IN_CI=true \
@@ -297,7 +255,7 @@ jobs:
         env:
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
 
-  # Run unit tests with the EventStream runtime Docker images as openhands user
+  # Run unit tests with the Docker runtime Docker images as openhands user
   test_runtime_oh:
     name: RT Unit Tests (openhands)
     runs-on: ubuntu-latest
@@ -307,20 +265,6 @@ jobs:
         base_image: ['nikolaik']
     steps:
       - uses: actions/checkout@v4
-      - name: Free Disk Space (Ubuntu)
-        uses: jlumbroso/free-disk-space@main
-        with:
-          # this might remove tools that are actually needed,
-          # if set to "true" but frees about 6 GB
-          tool-cache: true
-          # all of these default to true, but feel free to set to
-          # "false" if necessary for your workflow
-          android: true
-          dotnet: true
-          haskell: true
-          large-packages: true
-          docker-images: false
-          swap-storage: true
       - name: Set up Docker Buildx
         id: buildx
         uses: docker/setup-buildx-action@v3
@@ -363,7 +307,7 @@ jobs:
           image_name=ghcr.io/${{ github.repository_owner }}/runtime:${{ env.RELEVANT_SHA }}-${{ matrix.base_image }}
           image_name=$(echo $image_name | tr '[:upper:]' '[:lower:]')
 
-          TEST_RUNTIME=eventstream \
+          TEST_RUNTIME=docker \
           SANDBOX_USER_ID=$(id -u) \
           SANDBOX_RUNTIME_CONTAINER_IMAGE=$image_name \
           TEST_IN_CI=true \

.github/workflows/openhands-resolver.yml

@@ -20,6 +20,10 @@ on:
         required: false
         type: string
         default: "anthropic/claude-3-5-sonnet-20241022"
+      LLM_API_VERSION:
+        required: false
+        type: string
+        default: ""
       base_container_image:
         required: false
         type: string
@@ -84,12 +88,10 @@ jobs:
         run: |
           python -m pip index versions openhands-ai > openhands_versions.txt
           OPENHANDS_VERSION=$(head -n 1 openhands_versions.txt | awk '{print $2}' | tr -d '()')
-          # Ensure requirements.txt ends with newline before appending
-          if [ -f requirements.txt ] && [ -s requirements.txt ]; then
-            sed -i -e '$a\' requirements.txt
-          fi
-          echo "openhands-ai==${OPENHANDS_VERSION}" >> requirements.txt
-          cat requirements.txt
+
+          # Create a new requirements.txt locally within the workflow, ensuring no reference to the repo's file
+          echo "openhands-ai==${OPENHANDS_VERSION}" > /tmp/requirements.txt
+          cat /tmp/requirements.txt
 
       - name: Cache pip dependencies
         if: |
@@ -107,15 +109,16 @@ jobs:
         uses: actions/cache@v4
         with:
           path: ${{ env.pythonLocation }}/lib/python3.12/site-packages/*
-          key: ${{ runner.os }}-pip-openhands-resolver-${{ hashFiles('requirements.txt') }}
+          key: ${{ runner.os }}-pip-openhands-resolver-${{ hashFiles('/tmp/requirements.txt') }}
           restore-keys: |
-            ${{ runner.os }}-pip-openhands-resolver-${{ hashFiles('requirements.txt') }}
+            ${{ runner.os }}-pip-openhands-resolver-${{ hashFiles('/tmp/requirements.txt') }}
 
       - name: Check required environment variables
         env:
           LLM_MODEL: ${{ secrets.LLM_MODEL || inputs.LLM_MODEL }}
           LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
           LLM_BASE_URL: ${{ secrets.LLM_BASE_URL }}
+          LLM_API_VERSION: ${{ inputs.LLM_API_VERSION }}
           PAT_TOKEN: ${{ secrets.PAT_TOKEN }}
           PAT_USERNAME: ${{ secrets.PAT_USERNAME }}
           GITHUB_TOKEN: ${{ github.token }}
@@ -172,7 +175,7 @@ jobs:
           echo "SANDBOX_ENV_BASE_CONTAINER_IMAGE=${{ inputs.base_container_image }}" >> $GITHUB_ENV
 
           # Set branch variables
-          echo "TARGET_BRANCH=${{ inputs.target_branch }}" >> $GITHUB_ENV
+          echo "TARGET_BRANCH=${{ inputs.target_branch || 'main' }}" >> $GITHUB_ENV
 
       - name: Comment on issue with start message
         uses: actions/github-script@v7
@@ -220,16 +223,18 @@ jobs:
             } else {
               console.log("Installing from requirements.txt...");
               await exec.exec("python -m pip install --upgrade pip");
-              await exec.exec("pip install -r requirements.txt");
+              await exec.exec("pip install -r /tmp/requirements.txt");
             }
 
       - name: Attempt to resolve issue
         env:
           GITHUB_TOKEN: ${{ secrets.PAT_TOKEN || github.token }}
           GITHUB_USERNAME: ${{ secrets.PAT_USERNAME || 'openhands-agent' }}
+          GIT_USERNAME: ${{ secrets.PAT_USERNAME || 'openhands-agent' }}
           LLM_MODEL: ${{ secrets.LLM_MODEL || inputs.LLM_MODEL }}
           LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
           LLM_BASE_URL: ${{ secrets.LLM_BASE_URL }}
+          LLM_API_VERSION: ${{ inputs.LLM_API_VERSION }}
           PYTHONPATH: ""
         run: |
           cd /tmp && python -m openhands.resolver.resolve_issue \
@@ -262,14 +267,17 @@ jobs:
         env:
           GITHUB_TOKEN: ${{ secrets.PAT_TOKEN || github.token }}
           GITHUB_USERNAME: ${{ secrets.PAT_USERNAME || 'openhands-agent' }}
+          GIT_USERNAME: ${{ secrets.PAT_USERNAME || 'openhands-agent' }}
           LLM_MODEL: ${{ secrets.LLM_MODEL || inputs.LLM_MODEL }}
           LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
           LLM_BASE_URL: ${{ secrets.LLM_BASE_URL }}
+          LLM_API_VERSION: ${{ inputs.LLM_API_VERSION }}
           PYTHONPATH: ""
         run: |
           if [ "${{ steps.check_result.outputs.RESOLUTION_SUCCESS }}" == "true" ]; then
             cd /tmp && python -m openhands.resolver.send_pull_request \
               --issue-number ${{ env.ISSUE_NUMBER }} \
+              --target-branch ${{ env.TARGET_BRANCH }} \
               --pr-type draft \
               --reviewer ${{ github.actor }} | tee pr_result.txt && \
               grep "draft created" pr_result.txt | sed 's/.*\///g' > pr_number.txt

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

@@ -1,98 +0,0 @@
-# Workflow that runs python unit tests on mac
-name: Run Python Unit Tests Mac
-
-# This job is flaky so only run it nightly
-on:
-  schedule:
-    - cron: '0 0 * * *'
-
-jobs:
-  # Run python unit tests on macOS
-  test-on-macos:
-    name: Python Unit Tests on macOS
-    runs-on: macos-14
-    env:
-      INSTALL_DOCKER: '1' # Set to '0' to skip Docker installation
-    strategy:
-      matrix:
-        python-version: ['3.12']
-    steps:
-      - uses: actions/checkout@v4
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v5
-        with:
-          python-version: ${{ matrix.python-version }}
-      - name: Cache Poetry dependencies
-        uses: actions/cache@v4
-        with:
-          path: |
-            ~/.cache/pypoetry
-            ~/.virtualenvs
-          key: ${{ runner.os }}-poetry-${{ hashFiles('**/poetry.lock') }}
-          restore-keys: |
-            ${{ runner.os }}-poetry-
-      - name: Install tmux
-        run: brew install tmux
-      - name: Install poetry via pipx
-        run: pipx install poetry
-      - name: Install Python dependencies using Poetry
-        run: poetry install --without evaluation,llama-index
-      - name: Install & Start Docker
-        if: env.INSTALL_DOCKER == '1'
-        run: |
-          INSTANCE_NAME="colima-${GITHUB_RUN_ID}"
-
-          # Uninstall colima to upgrade to the latest version
-          if brew list colima &>/dev/null; then
-            brew uninstall colima
-            # unlinking colima dependency: go
-            brew uninstall go@1.21
-          fi
-          rm -rf ~/.colima ~/.lima
-          brew install --HEAD colima
-          brew install docker
-
-          start_colima() {
-            # Find a free port in the range 10000-20000
-            RANDOM_PORT=$((RANDOM % 10001 + 10000))
-
-            # Original line:
-            if ! colima start --network-address --arch x86_64 --cpu=1 --memory=1 --verbose --ssh-port $RANDOM_PORT; then
-              echo "Failed to start Colima."
-              return 1
-            fi
-            return 0
-          }
-
-          # Attempt to start Colima for 5 total attempts:
-          ATTEMPT_LIMIT=5
-          for ((i=1; i<=ATTEMPT_LIMIT; i++)); do
-
-            if start_colima; then
-              echo "Colima started successfully."
-              break
-            else
-              colima stop -f
-              sleep 10
-              colima delete -f
-              if [ $i -eq $ATTEMPT_LIMIT ]; then
-                exit 1
-              fi
-              sleep 10
-            fi
-          done
-
-          # For testcontainers to find the Colima socket
-          # https://github.com/abiosoft/colima/blob/main/docs/FAQ.md#cannot-connect-to-the-docker-daemon-at-unixvarrundockersock-is-the-docker-daemon-running
-          sudo ln -sf $HOME/.colima/default/docker.sock /var/run/docker.sock
-      - name: Build Environment
-        run: make build
-      - name: Set up Docker Buildx
-        id: buildx
-        uses: docker/setup-buildx-action@v3
-      - name: Run Tests
-        run: poetry run pytest --forked --cov=openhands --cov-report=xml ./tests/unit --ignore=tests/unit/test_memory.py
-      - name: Upload coverage to Codecov
-        uses: codecov/codecov-action@v5
-        env:
-          CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}

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

@@ -32,6 +32,10 @@ jobs:
         uses: docker/setup-buildx-action@v3
       - name: Install tmux
         run: sudo apt-get update && sudo apt-get install -y tmux
+      - name: Setup Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: '22.x'
       - name: Install poetry via pipx
         run: pipx install poetry
       - name: Set up Python
@@ -44,7 +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:

.github/workflows/stale.yml

@@ -19,3 +19,4 @@ jobs:
           close-issue-message: 'This issue was closed because it has been stalled for over 30 days with no activity.'
           close-pr-message: 'This PR was closed because it has been stalled for over 30 days with no activity.'
           days-before-close: 7
+          operations-per-run: 150

.openhands/microagents/glossary.md

@@ -0,0 +1,172 @@
+# OpenHands Glossary
+
+### Agent
+The core AI entity in OpenHands that can perform software development tasks by interacting with tools, browsing the web, and modifying code.
+
+#### Agent Controller
+A component that manages the agent's lifecycle, handles its state, and coordinates interactions between the agent and various tools.
+
+#### Agent Delegation
+The ability of an agent to hand off specific tasks to other specialized agents for better task completion.
+
+#### Agent Hub
+A central registry of different agent types and their capabilities, allowing for easy agent selection and instantiation.
+
+#### Agent Skill
+A specific capability or function that an agent can perform, such as file manipulation, web browsing, or code editing.
+
+#### Agent State
+The current context and status of an agent, including its memory, active tools, and ongoing tasks.
+
+#### CodeAct Agent
+[A generalist agent in OpenHands](https://arxiv.org/abs/2407.16741) designed to perform tasks by editing and executing code.
+
+### Browser
+A system for web-based interactions and tasks.
+
+#### Browser Gym
+A testing and evaluation environment for browser-based agent interactions and tasks.
+
+#### Web Browser Tool
+A tool that enables agents to interact with web pages and perform web-based tasks.
+
+### Commands
+Terminal and execution related functionality.
+
+#### Bash Session
+A persistent terminal session that maintains state and history for bash command execution.
+This uses tmux under the hood.
+
+### Configuration
+System-wide settings and options.
+
+#### Agent Configuration
+Settings that define an agent's behavior, capabilities, and limitations, including available tools and runtime settings.
+
+#### Configuration Options
+Settings that control various aspects of OpenHands behavior, including runtime, security, and agent settings.
+
+#### LLM Config
+Configuration settings for language models used by agents, including model selection and parameters.
+
+#### LLM Draft Config
+Settings for draft mode operations with language models, typically used for faster, lower-quality responses.
+
+#### Runtime Configuration
+Settings that define how the runtime environment should be set up and operated.
+
+#### Security Options
+Configuration settings that control security features and restrictions.
+
+### Conversation
+A sequence of interactions between a user and an agent, including messages, actions, and their results.
+
+#### Conversation Info
+Metadata about a conversation, including its status, participants, and timeline.
+
+#### Conversation Manager
+A component that handles the creation, storage, and retrieval of conversations.
+
+#### Conversation Metadata
+Additional information about conversations, such as tags, timestamps, and related resources.
+
+#### Conversation Status
+The current state of a conversation, including whether it's active, completed, or failed.
+
+#### Conversation Store
+A storage system for maintaining conversation history and related data.
+
+### Events
+
+#### Event
+Every Conversation comprises a series of Events. Each Event is either an Action or an Observation.
+
+#### Event Stream
+A continuous flow of events that represents the ongoing activities and interactions in the system.
+
+#### Action
+A specific operation or command that an agent executes through available tools, such as running a command or editing a file.
+
+#### Observation
+The response or result returned by a tool after an agent's action, providing feedback about the action's outcome.
+
+### Interface
+Different ways to interact with OpenHands.
+
+#### CLI Mode
+A command-line interface mode for interacting with OpenHands agents without a graphical interface.
+
+#### GUI Mode
+A graphical user interface mode for interacting with OpenHands agents through a web interface.
+
+#### Headless Mode
+A mode of operation where OpenHands runs without a user interface, suitable for automation and scripting.
+
+### Agent Memory
+The system that decides which parts of the Event Stream (i.e. the conversation history) should be passed into each LLM prompt.
+
+#### Memory Store
+A storage system for maintaining agent memory and context across sessions.
+
+#### Condenser
+A component that processes and summarizes conversation history to maintain context while staying within token limits.
+
+#### Truncation
+A very simple Condenser strategy. Reduces conversation history or content to stay within token limits.
+
+### Microagent
+A specialized prompt that enhances OpenHands with domain-specific knowledge, repository-specific context, and task-specific workflows.
+
+#### Microagent Registry
+A central repository of available microagents and their configurations.
+
+#### Public Microagent
+A general-purpose microagent available to all OpenHands users, triggered by specific keywords.
+
+#### Repository Microagent
+A type of microagent that provides repository-specific context and guidelines, stored in the `.openhands/microagents/` directory.
+
+### Prompt
+Components for managing and processing prompts.
+
+#### Prompt Caching
+A system for caching and reusing common prompts to improve performance.
+
+#### Prompt Manager
+A component that handles the loading, processing, and management of prompts used by agents, including microagents.
+
+#### Response Parsing
+The process of interpreting and structuring responses from language models and tools.
+
+### Runtime
+The execution environment where agents perform their tasks, which can be local, remote, or containerized.
+
+#### Action Execution Server
+A REST API that receives agent actions (e.g. bash commands, python code, browsing actions), executes them in the runtime environment, and returns the results.
+
+#### Action Execution Client
+A component that handles the execution of actions in the runtime environment, managing the communication between the agent and the runtime.
+
+#### Docker Runtime
+A containerized runtime environment that provides isolation and reproducibility for agent operations.
+
+#### E2B Runtime
+A specialized runtime environment built on E2B for secure and isolated code execution.
+
+#### Local Runtime
+A runtime environment that executes on the local machine, suitable for development and testing.
+
+#### Modal Runtime
+A runtime environment built on Modal for scalable and distributed agent operations.
+
+#### Remote Runtime
+A sandboxed environment that executes code and commands remotely, providing isolation and security for agent operations.
+
+#### Runtime Builder
+A component that builds a Docker image for the Action Execution Server based on a user-specified base image.
+
+### Security
+Security-related components and features.
+
+#### Security Analyzer
+A component that checks agent actions for potential security risks.

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:0.21-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
@@ -81,10 +81,10 @@ check-nodejs:
 	@if command -v node > /dev/null; then \
 		NODE_VERSION=$(shell node --version | sed -E 's/v//g'); \
 		IFS='.' read -r -a NODE_VERSION_ARRAY <<< "$$NODE_VERSION"; \
-		if [ "$${NODE_VERSION_ARRAY[0]}" -ge 20 ]; then \
+		if [ "$${NODE_VERSION_ARRAY[0]}" -ge 22 ]; then \
 			echo "$(BLUE)Node.js $$NODE_VERSION is already installed.$(RESET)"; \
 		else \
-			echo "$(RED)Node.js 20.x or later is required. Please install Node.js 20.x or later to continue.$(RESET)"; \
+			echo "$(RED)Node.js 22.x or later is required. Please install Node.js 22.x or later to continue.$(RESET)"; \
 			exit 1; \
 		fi; \
 	else \

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:0.21-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:0.21-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:0.21
+    docker.all-hands.dev/all-hands-ai/openhands:0.26
 ```
 
 You'll find OpenHands running at [http://localhost:3000](http://localhost:3000)!

build.sh

@@ -1,5 +1,4 @@
-#!/bin/bash
+#!/usr/bin/env bash
 set -e
 
-cp pyproject.toml poetry.lock openhands
 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"
 
@@ -104,7 +110,7 @@ workspace_base = "./workspace"
 #aws_secret_access_key = ""
 
 # API key to use (For Headless / CLI only -  In Web this is overridden by Session Init)
-api_key = "your-api-key"
+api_key = ""
 
 # API base URL (For Headless / CLI only -  In Web this is overridden by Session Init)
 #base_url = ""
@@ -195,7 +201,7 @@ model = "gpt-4o"
 #native_tool_calling = None
 
 [llm.gpt4o-mini]
-api_key = "your-api-key"
+api_key = ""
 model = "gpt-4o"
 
 
@@ -234,6 +240,10 @@ codeact_enable_jupyter = true
 # List of microagents to disable
 #disabled_microagents = []
 
+# Whether history should be truncated to continue the session when hitting LLM context
+# length limit
+enable_history_truncation = true
+
 [agent.RepoExplorerAgent]
 # Example: use a cheaper model for RepoExplorerAgent to reduce cost, especially
 # useful when an agent doesn't demand high quality but uses a lot of tokens

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:0.21-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() {

dev_config/python/ruff.toml

@@ -24,3 +24,6 @@ inline-quotes = "single"
 
 [format]
 quote-style = "single"
+
+[lint.flake8-bugbear]
+extend-immutable-calls = ["Depends", "fastapi.Depends", "fastapi.params.Depends"]

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:0.21-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/DOC_STYLE_GUIDE.md

@@ -46,3 +46,11 @@ docker run -it \
     -e THAT=that
     ...
 ```
+
+### Referring to UI Elements
+
+When referencing UI elements, use ``.
+
+Example:
+1. Toggle the `Advanced` option
+2. Enter your model in the `Custom Model` textbox.

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

@@ -1,8 +1,8 @@
 
 
-# 📦 Runtime EventStream
+# 📦 Runtime Docker
 
-Le Runtime EventStream d'OpenHands est le composant principal qui permet l'exécution sécurisée et flexible des actions des agents d'IA.
+Le Runtime Docker d'OpenHands est le composant principal qui permet l'exécution sécurisée et flexible des actions des agents d'IA.
 Il crée un environnement en bac à sable (sandbox) en utilisant Docker, où du code arbitraire peut être exécuté en toute sécurité sans risquer le système hôte.
 
 ## Pourquoi avons-nous besoin d'un runtime en bac à sable ?

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

@@ -163,7 +163,7 @@ Les options de configuration de base sont définies dans la section `[core]` du
 
 - `runtime`
   - Type : `str`
-  - Valeur par défaut : `"eventstream"`
+  - Valeur par défaut : `"docker"`
   - Description : Environnement d'exécution
 
 - `default_agent`

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:0.21-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:0.21 \
+    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/evaluation-harness.md

@@ -114,7 +114,7 @@ Pour créer un workflow d'évaluation pour votre benchmark, suivez ces étapes :
    def get_config(instance: pd.Series, metadata: EvalMetadata) -> AppConfig:
        config = AppConfig(
            default_agent=metadata.agent_class,
-           runtime='eventstream',
+           runtime='docker',
            max_iterations=metadata.max_iterations,
            sandbox=SandboxConfig(
                base_container_image='your_container_image',

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:0.21-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:0.21 \
+    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:0.21-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:0.21-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:0.21
+    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:0.21-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/architecture/runtime.md

@@ -1,8 +1,8 @@
 以下是翻译后的内容:
 
-# 📦 EventStream 运行时
+# 📦 Docker 运行时
 
-OpenHands EventStream 运行时是实现 AI 代理操作安全灵活执行的核心组件。
+OpenHands Docker 运行时是实现 AI 代理操作安全灵活执行的核心组件。
 它使用 Docker 创建一个沙盒环境,可以安全地运行任意代码而不会危及主机系统。
 
 ## 为什么我们需要沙盒运行时?

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

@@ -162,7 +162,7 @@
 
 - `runtime`
   - 类型: `str`
-  - 默认值: `"eventstream"`
+  - 默认值: `"docker"`
   - 描述: 运行时环境
 
 - `default_agent`

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:0.21-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:0.21 \
+    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/evaluation-harness.md

@@ -112,7 +112,7 @@ OpenHands 的主要入口点在 `openhands/core/main.py` 中。以下是它的
    def get_config(instance: pd.Series, metadata: EvalMetadata) -> AppConfig:
        config = AppConfig(
            default_agent=metadata.agent_class,
-           runtime='eventstream',
+           runtime='docker',
            max_iterations=metadata.max_iterations,
            sandbox=SandboxConfig(
                base_container_image='your_container_image',

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:0.21-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:0.21 \
+    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:0.21-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:0.21-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:0.21
+    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:0.21-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/architecture/runtime.md

@@ -1,6 +1,6 @@
-# 📦 EventStream Runtime
+# 📦 Docker Runtime
 
-The OpenHands EventStream Runtime is the core component that enables secure and flexible execution of AI agent's action.
+The OpenHands Docker Runtime is the core component that enables secure and flexible execution of AI agent's action.
 It creates a sandboxed environment using Docker, where arbitrary code can be run safely without risking the host system.
 
 ## Why do we need a sandboxed runtime?
@@ -54,14 +54,13 @@ graph TD
 6. Action Execution: The runtime client receives actions from the backend, executes them in the sandboxed environment, and sends back observations
 7. Observation Return: The action execution server sends execution results back to the OpenHands backend as observations
 
-
 The role of the client:
+
 - It acts as an intermediary between the OpenHands backend and the sandboxed environment
 - It executes various types of actions (shell commands, file operations, Python code, etc.) safely within the container
 - It manages the state of the sandboxed environment, including the current working directory and loaded plugins
 - It formats and returns observations to the backend, ensuring a consistent interface for processing results
 
-
 ## How OpenHands builds and maintains OH Runtime images
 
 OpenHands' approach to building and managing runtime images ensures efficiency, consistency, and flexibility in creating and maintaining Docker images for both production and development environments.
@@ -78,16 +77,15 @@ Tags may be in one of 2 formats:
 - **Source Tag**: `oh_v{openhands_version}_{16_digit_lock_hash}_{16_digit_source_hash}`
   (e.g.: `oh_v0.9.9_1234567890abcdef_1234567890abcdef`)
 
-
 #### Source Tag - Most Specific
 
 This is the first 16 digits of the MD5 of the directory hash for the source directory. This gives a hash
 for only the openhands source
 
-
 #### Lock Tag
 
 This hash is built from the first 16 digits of the MD5 of:
+
 - The name of the base image upon which the image was built (e.g.: `nikolaik/python-nodejs:python3.12-nodejs22`)
 - The content of the `pyproject.toml` included in the image.
 - The content of the `poetry.lock` included in the image.

docs/modules/usage/cloud/cloud-github-resolver.md

@@ -0,0 +1,21 @@
+# Cloud GitHub Resolver
+
+The GitHub Resolver automates code fixes and provides intelligent assistance for your repositories.
+
+## Setup
+
+The Cloud Github Resolver is available automatically when you
+[grant OpenHands Cloud repository access](./openhands-cloud.md#adding-repositories).
+
+## Usage
+
+### Issues
+
+On your repository, label an issue with `openhands`. OpenHands will attempt to fix the issue.
+
+### Pull Requests
+
+In order to get OpenHands to work on pull requests, use `@openhands` in top level or single inline comments to:
+     - Ask questions
+     - Request updates
+     - Get code explanations

docs/modules/usage/cloud/openhands-cloud.md

@@ -0,0 +1,39 @@
+# Openhands Cloud
+
+OpenHands Cloud is the cloud hosted version of OpenHands by All Hands AI.
+
+## Accessing OpenHands Cloud
+
+Currently, users are being admitted to access OpenHands Cloud in waves. To sign up,
+[join the waitlist](https://www.all-hands.dev/join-waitlist). Once you are approved, you will get an email with
+instructions on how to access it.
+
+## Getting Started
+
+After visiting OpenHands Cloud, you will be asked to connect with your GitHub account:
+1. After reading and accepting the terms of service, click `Connect to GitHub`.
+2. Review the permissions requested by OpenHands and then click `Authorize OpenHands by All Hands AI`.
+   - OpenHands will require some permissions from your GitHub account. To read more about these permissions,
+     you can click the `Learn more` link on the GitHub authorize page.
+
+## Adding Repositories
+
+You can grant OpenHands specific repository access:
+1. Under the `Select a GitHub project` dropdown, select `Add more repositories...`.
+2. Select the organization, then choose the specific repositories to grant OpenHands access to.
+   - Openhands requests short-lived tokens (8-hour expiry) with these permissions:
+     - Actions: Read and write
+     - Administration: Read-only
+     - Commit statuses: Read and write
+     - Contents: Read and write
+     - Issues: Read and write
+     - Metadata: Read-only
+     - Pull requests: Read and write
+     - Webhooks: Read and write
+     - Workflows: Read and write
+   - Repository access for a user is granted based on:
+     - Granted permission for the repository.
+     - User's GitHub permissions (owner/collaborator).
+
+You can manage repository access any time by following the above workflow or visiting the Settings page and selecting
+`Configure GitHub Repositories` under the `GitHub Settings` section.

docs/modules/usage/configuration-options.md

@@ -126,7 +126,7 @@ The core configuration options are defined in the `[core]` section of the `confi
 
 - `runtime`
   - Type: `str`
-  - Default: `"eventstream"`
+  - Default: `"docker"`
   - Description: Runtime environment
 
 - `default_agent`
@@ -340,6 +340,11 @@ The agent configuration options are defined in the `[agent]` and `[agent.<agent_
   - Default: `false`
   - Description: Whether Jupyter is enabled in the action space
 
+- `enable_history_truncation`
+  - Type: `bool`
+  - Default: `true`
+  - Description: Whether history should be truncated to continue the session when hitting LLM context length limit
+
 ### Microagent Usage
 - `enable_prompt_extensions`
   - Type: `bool`

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:0.21-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:0.21 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.26 \
     python -m openhands.core.cli
 ```
 

docs/modules/usage/how-to/custom-sandbox-guide.md

@@ -41,8 +41,16 @@ docker build -t custom-image .
 
 This will produce a new image called `custom-image`, which will be available in Docker.
 
-> Note that in the configuration described in this document, OpenHands will run as user "openhands" inside the
-> sandbox and thus all packages installed via the docker file should be available to all users on the system, not just root.
+## Using the Docker Command
+
+When running OpenHands using [the docker command](/modules/usage/installation#start-the-app), replace
+`-e SANDBOX_RUNTIME_CONTAINER_IMAGE=...` with `-e SANDBOX_BASE_CONTAINER_IMAGE=<custom image name>`:
+
+```commandline
+docker run -it --rm --pull=always \
+    -e SANDBOX_BASE_CONTAINER_IMAGE=custom-image \
+    ...
+```
 
 ## Using the Development Workflow
 

docs/modules/usage/how-to/evaluation-harness.md

@@ -112,7 +112,7 @@ To create an evaluation workflow for your benchmark, follow these steps:
    def get_config(instance: pd.Series, metadata: EvalMetadata) -> AppConfig:
        config = AppConfig(
            default_agent=metadata.agent_class,
-           runtime='eventstream',
+           runtime='docker',
            max_iterations=metadata.max_iterations,
            sandbox=SandboxConfig(
                base_container_image='your_container_image',

docs/modules/usage/how-to/github-action.md

@@ -42,9 +42,10 @@ You can provide custom directions for OpenHands by following the [README for the
 Github resolver will automatically check for valid [repository secrets](https://docs.github.com/en/actions/security-for-github-actions/security-guides/using-secrets-in-github-actions?tool=webui#creating-secrets-for-a-repository) or [repository variables](https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/store-information-in-variables#creating-configuration-variables-for-a-repository) to customize its behavior.
 The customization options you can set are:
 
-| **Attribute name**               | **Type** | **Purpose**                                                                                                 | **Example**                                          |
-|----------------------------------| -------- |-------------------------------------------------------------------------------------------------------------|------------------------------------------------------|
-| `LLM_MODEL`                      | Variable | Set the LLM to use with OpenHands                                                                           | `LLM_MODEL="anthropic/claude-3-5-sonnet-20241022"`   |
-| `OPENHANDS_MAX_ITER`             | Variable | Set max limit for agent iterations                                                                          | `OPENHANDS_MAX_ITER=10`                              |
-| `OPENHANDS_MACRO`                | Variable | Customize default macro for invoking the resolver                                                           | `OPENHANDS_MACRO=@resolveit`                         |
-| `OPENHANDS_BASE_CONTAINER_IMAGE` | Variable | Custom Sandbox ([learn more](https://docs.all-hands.dev/modules/usage/how-to/custom-sandbox-guide))         | `OPENHANDS_BASE_CONTAINER_IMAGE="custom_image"`      |
+| **Attribute name**               | **Type** | **Purpose**                                                                                         | **Example**                                        |
+| -------------------------------- | -------- | --------------------------------------------------------------------------------------------------- | -------------------------------------------------- |
+| `LLM_MODEL`                      | Variable | Set the LLM to use with OpenHands                                                                   | `LLM_MODEL="anthropic/claude-3-5-sonnet-20241022"` |
+| `OPENHANDS_MAX_ITER`             | Variable | Set max limit for agent iterations                                                                  | `OPENHANDS_MAX_ITER=10`                            |
+| `OPENHANDS_MACRO`                | Variable | Customize default macro for invoking the resolver                                                   | `OPENHANDS_MACRO=@resolveit`                       |
+| `OPENHANDS_BASE_CONTAINER_IMAGE` | Variable | Custom Sandbox ([learn more](https://docs.all-hands.dev/modules/usage/how-to/custom-sandbox-guide)) | `OPENHANDS_BASE_CONTAINER_IMAGE="custom_image"`    |
+| `TARGET_BRANCH`                  | Variable | Merge to branch other than `main`                                                                   | `TARGET_BRANCH="dev"`                              |

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

@@ -1,9 +1,6 @@
 # GUI Mode
 
-## Introduction
-
-OpenHands provides a user-friendly Graphical User Interface (GUI) mode for interacting with the AI assistant.
-This mode offers an intuitive way to set up the environment, manage settings, and communicate with the AI.
+OpenHands provides a Graphical User Interface (GUI) mode for interacting with the AI assistant.
 
 ## Installation and Setup
 
@@ -14,104 +11,95 @@ This mode offers an intuitive way to set up the environment, manage settings, an
 
 ### Initial Setup
 
-1. Upon first launch, you'll see a settings modal.
-2. Select an `LLM Provider` and `LLM Model` from the dropdown menus.
+1. Upon first launch, you'll see a settings page.
+2. Select an `LLM Provider` and `LLM Model` from the dropdown menus. If the required model does not exist in the list,
+   toggle `Advanced` options and enter it with the correct prefix in the `Custom Model` text box.
 3. Enter the corresponding `API Key` for your chosen provider.
-4. Click "Save" to apply the settings.
+4. Click `Save Changes` to apply the settings.
 
 ### GitHub Token Setup
 
 OpenHands automatically exports a `GITHUB_TOKEN` to the shell environment if it is available. This can happen in two ways:
 
-- **Locally (OSS)**: The user directly inputs their GitHub token.
-- **Online (SaaS)**: The token is obtained through GitHub OAuth authentication.
-
-#### Setting Up a Local GitHub Token
-
-1. **Generate a Personal Access Token (PAT)**:
-   - Go to GitHub Settings > Developer Settings > Personal Access Tokens > Tokens (classic).
-   - Click "Generate new token (classic)".
+- **Local Installation**: The user directly inputs their GitHub token.
+<details>
+  <summary>Setting Up a GitHub Token</summary>
+  1. **Generate a Personal Access Token (PAT)**:
+   - On GitHub, go to Settings > Developer Settings > Personal Access Tokens > Tokens (classic).
+   - Click `Generate new token (classic)`.
    - Required scopes:
      - `repo` (Full control of private repositories)
-     - `workflow` (Update GitHub Action workflows)
-     - `read:org` (Read organization data)
+  2. **Enter Token in OpenHands**:
+   - Click the Settings button (gear icon).
+   - Navigate to the `GitHub Settings` section.
+   - Paste your token in the `GitHub Token` field.
+   - Click `Save Changes` to apply the changes.
+</details>
 
-2. **Enter Token in OpenHands**:
-   - Click the Settings button (gear icon) in the top right.
-   - Navigate to the "GitHub" section.
-   - Paste your token in the "GitHub Token" field.
-   - Click "Save" to apply the changes.
+<details>
+  <summary>Organizational Token Policies</summary>
 
-#### Organizational Token Policies
+  If you're working with organizational repositories, additional setup may be required:
 
-If you're working with organizational repositories, additional setup may be required:
-
-1. **Check Organization Requirements**:
+  1. **Check Organization Requirements**:
    - Organization admins may enforce specific token policies.
    - Some organizations require tokens to be created with SSO enabled.
    - Review your organization's [token policy settings](https://docs.github.com/en/organizations/managing-programmatic-access-to-your-organization/setting-a-personal-access-token-policy-for-your-organization).
-
-2. **Verify Organization Access**:
+  2. **Verify Organization Access**:
    - Go to your token settings on GitHub.
-   - Look for the organization under "Organization access".
-   - If required, click "Enable SSO" next to your organization.
+   - Look for the organization under `Organization access`.
+   - If required, click `Enable SSO` next to your organization.
    - Complete the SSO authorization process.
+</details>
 
-#### OAuth Authentication (Online Mode)
+<details>
+  <summary>Troubleshooting</summary>
 
-When using OpenHands in online mode, the GitHub OAuth flow:
+  Common issues and solutions:
 
-1. Requests the following permissions:
+  - **Token Not Recognized**:
+     - Ensure the token is properly saved in settings.
+     - Check that the token hasn't expired.
+     - Verify the token has the required scopes.
+     - Try regenerating the token.
+
+  - **Organization Access Denied**:
+     - Check if SSO is required but not enabled.
+     - Verify organization membership.
+     - Contact organization admin if token policies are blocking access.
+
+  - **Verifying Token Works**:
+     - The app will show a green checkmark if the token is valid.
+     - Try accessing a repository to confirm permissions.
+     - Check the browser console for any error messages.
+</details>
+
+- **OpenHands Cloud**: The token is obtained through GitHub OAuth authentication.
+
+<details>
+  <summary>OAuth Authentication</summary>
+
+  When using OpenHands Cloud, the GitHub OAuth flow requests the following permissions:
    - Repository access (read/write)
    - Workflow management
    - Organization read access
 
-2. Authentication steps:
-   - Click "Sign in with GitHub" when prompted.
+  To authenticate OpenHands:
+   - Click `Sign in with GitHub` when prompted.
    - Review the requested permissions.
    - Authorize OpenHands to access your GitHub account.
    - If using an organization, authorize organization access if prompted.
-
-#### Troubleshooting
-
-Common issues and solutions:
-
-- **Token Not Recognized**:
-   - Ensure the token is properly saved in settings.
-   - Check that the token hasn't expired.
-   - Verify the token has the required scopes.
-   - Try regenerating the token.
-
-- **Organization Access Denied**:
-   - Check if SSO is required but not enabled.
-   - Verify organization membership.
-   - Contact organization admin if token policies are blocking access.
-
-- **Verifying Token Works**:
-   - The app will show a green checkmark if the token is valid.
-   - Try accessing a repository to confirm permissions.
-   - Check the browser console for any error messages.
-   - Use the "Test Connection" button in settings if available.
+</details>
 
 ### Advanced Settings
 
-1. Toggle `Advanced Options` to access additional settings.
+1. Inside the Settings page, toggle `Advanced` options to access additional settings.
 2. Use the `Custom Model` text box to manually enter a model if it's not in the list.
 3. Specify a `Base URL` if required by your LLM provider.
 
-### Main Interface
-
-The main interface consists of several key components:
-
-- **Chat Window**: The central area where you can view the conversation history with the AI assistant.
-- **Input Box**: Located at the bottom of the screen, use this to type your messages or commands to the AI.
-- **Send Button**: Click this to send your message to the AI.
-- **Settings Button**: A gear icon that opens the settings modal, allowing you to adjust your configuration at any time.
-- **Workspace Panel**: Displays the files and folders in your workspace, allowing you to navigate and view files, or the agent's past commands or web browsing history.
-
 ### Interacting with the AI
 
-1. Type your question, request, or task description in the input box.
+1. Type your prompt in the input box.
 2. Click the send button or press Enter to submit your message.
 3. The AI will process your input and provide a response in the chat window.
 4. You can continue the conversation by asking follow-up questions or providing additional information.

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:0.21-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:0.21 \
+    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

@@ -6,11 +6,14 @@
 - Linux
 - Windows with [WSL](https://learn.microsoft.com/en-us/windows/wsl/install) and [Docker Desktop support](https://docs.docker.com/desktop/setup/install/windows-install/#system-requirements)
 
+A system with a modern processor and a minimum of **4GB RAM** is recommended to run OpenHands.
+
 ## Prerequisites
 
 <details>
   <summary>MacOS</summary>
-  ### Docker Desktop
+
+  **Docker Desktop**
 
   1. [Install Docker Desktop on Mac](https://docs.docker.com/desktop/setup/install/mac-install).
   2. Open Docker Desktop, go to `Settings > Advanced` and ensure `Allow the default Docker socket to be used` is enabled.
@@ -23,7 +26,7 @@
   Tested with Ubuntu 22.04.
   :::
 
-  ### Docker Desktop
+  **Docker Desktop**
 
   1. [Install Docker Desktop on Linux](https://docs.docker.com/desktop/setup/install/linux/).
 
@@ -31,18 +34,23 @@
 
 <details>
   <summary>Windows</summary>
-  ### WSL
+
+  **WSL**
 
   1. [Install WSL](https://learn.microsoft.com/en-us/windows/wsl/install).
   2. Run `wsl --version` in powershell and confirm `Default Version: 2`.
 
-  ### Docker Desktop
+  **Docker Desktop**
 
   1. [Install Docker Desktop on Windows](https://docs.docker.com/desktop/setup/install/windows-install).
   2. Open Docker Desktop, go to `Settings` and confirm the following:
   - General: `Use the WSL 2 based engine` is enabled.
   - Resources > WSL Integration: `Enable integration with my default WSL distro` is enabled.
 
+  :::note
+  The docker command below to start the app must be run inside the WSL terminal.
+  :::
+
 </details>
 
 ## Start the App
@@ -50,17 +58,17 @@
 The easiest way to run OpenHands is in Docker.
 
 ```bash
-docker pull docker.all-hands.dev/all-hands-ai/runtime:0.21-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:0.21-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:0.21
+    docker.all-hands.dev/all-hands-ai/openhands:0.26
 ```
 
 You'll find OpenHands running at http://localhost:3000!
@@ -72,24 +80,22 @@ or run it on tagged issues with [a github action](https://docs.all-hands.dev/mod
 
 ## Setup
 
-Upon launching OpenHands, you'll see a settings modal. You **must** select an `LLM Provider` and `LLM Model` and enter a corresponding `API Key`.
+Upon launching OpenHands, you'll see a Settings page. You **must** select an `LLM Provider` and `LLM Model` and enter a corresponding `API Key`.
 These can be changed at any time by selecting the `Settings` button (gear icon) in the UI.
 
-If the required `LLM Model` does not exist in the list, you can toggle `Advanced Options` and manually enter it with the correct prefix
+If the required model does not exist in the list, you can toggle `Advanced` options and manually enter it with the correct prefix
 in the `Custom Model` text box.
-The `Advanced Options` also allow you to specify a `Base URL` if required.
+The `Advanced` options also allow you to specify a `Base URL` if required.
 
-<div style={{ display: 'flex', justifyContent: 'center', gap: '20px' }}>
-  <img src="/img/settings-screenshot.png" alt="settings-modal" width="340" />
-  <img src="/img/settings-advanced.png" alt="settings-modal" width="335" />
-</div>
+Now you're ready to [get started with OpenHands](./getting-started).
 
 ## Versions
 
-The command above pulls the most recent stable release of OpenHands. You have other options as well:
-- For a specific release, use `docker.all-hands.dev/all-hands-ai/openhands:$VERSION`, replacing $VERSION with the version number.
-- We use semver, and release major, minor, and patch tags. So `0.9` will automatically point to the latest `0.9.x` release, and `0` will point to the latest `0.x.x` release.
-- For the most up-to-date development version, you can use `docker.all-hands.dev/all-hands-ai/openhands:main`. This version is unstable and is recommended for testing or development purposes only.
+The [docker command above](./installation#start-the-app) pulls the most recent stable release of OpenHands. You have other options as well:
+- For a specific release, replace $VERSION in `openhands:$VERSION` and `runtime:$VERSION`, with the version number.
+We use SemVer so `0.9` will automatically point to the latest `0.9.x` release, and `0` will point to the latest `0.x.x` release.
+- For the most up-to-date development version, replace $VERSION in `openhands:$VERSION` and `runtime:$VERSION`, with `main`.
+This version is unstable and is recommended for testing or development purposes only.
 
 You can choose the tag that best suits your needs based on stability requirements and desired features.
 

docs/modules/usage/llms/azure-llms.md

@@ -25,7 +25,7 @@ You will need your ChatGPT deployment name which can be found on the deployments
 <deployment-name> below.
 :::
 
-1. Enable `Advanced Options`
+1. Enable `Advanced` options
 2. Set the following:
    - `Custom Model` to azure/<deployment-name>
    - `Base URL` to your Azure API Base URL (e.g. `https://example-endpoint.openai.azure.com`)

docs/modules/usage/llms/google-llms.md

@@ -10,7 +10,7 @@ OpenHands uses LiteLLM to make calls to Google's chat models. You can find their
 When running OpenHands, you'll need to set the following in the OpenHands UI through the Settings:
 - `LLM Provider` to `Gemini`
 - `LLM Model` to the model you will be using.
-If the model is not in the list, toggle `Advanced Options`, and enter it in `Custom Model` (e.g. gemini/<model-name> like `gemini/gemini-1.5-pro`).
+If the model is not in the list, toggle `Advanced` options, and enter it in `Custom Model` (e.g. gemini/<model-name> like `gemini/gemini-1.5-pro`).
 - `API Key` to your Gemini API key
 
 ## VertexAI - Google Cloud Platform Configs
@@ -27,4 +27,4 @@ VERTEXAI_LOCATION="<your-gcp-location>"
 Then set the following in the OpenHands UI through the Settings:
 - `LLM Provider` to `VertexAI`
 - `LLM Model` to the model you will be using.
-If the model is not in the list, toggle `Advanced Options`, and enter it in `Custom Model` (e.g. vertex_ai/&lt;model-name&gt;).
+If the model is not in the list, toggle `Advanced` options, and enter it in `Custom Model` (e.g. vertex_ai/&lt;model-name&gt;).

docs/modules/usage/llms/groq.md

@@ -8,7 +8,7 @@ When running OpenHands, you'll need to set the following in the OpenHands UI thr
 - `LLM Provider` to `Groq`
 - `LLM Model` to the model you will be using. [Visit here to see the list of
 models that Groq hosts](https://console.groq.com/docs/models). If the model is not in the list, toggle
-`Advanced Options`, and enter it in `Custom Model` (e.g. groq/<model-name> like `groq/llama3-70b-8192`).
+`Advanced` options, and enter it in `Custom Model` (e.g. groq/<model-name> like `groq/llama3-70b-8192`).
 - `API key` to your Groq API key. To find or create your Groq API Key, [see here](https://console.groq.com/keys).
 
 
@@ -17,7 +17,7 @@ models that Groq hosts](https://console.groq.com/docs/models). If the model is n
 
 The Groq endpoint for chat completion is [mostly OpenAI-compatible](https://console.groq.com/docs/openai). Therefore, you can access Groq models as you
 would access any OpenAI-compatible endpoint. In the OpenHands UI through the Settings:
-1. Enable `Advanced Options`
+1. Enable `Advanced` options
 2. Set the following:
    - `Custom Model` to the prefix `openai/` + the model you will be using (e.g. `openai/llama3-70b-8192`)
    - `Base URL` to `https://api.groq.com/openai/v1`

docs/modules/usage/llms/litellm-proxy.md

@@ -8,7 +8,7 @@ To use LiteLLM proxy with OpenHands, you need to:
 
 1. Set up a LiteLLM proxy server (see [LiteLLM documentation](https://docs.litellm.ai/docs/proxy/quick_start))
 2. When running OpenHands, you'll need to set the following in the OpenHands UI through the Settings:
-  * Enable `Advanced Options`
+  * Enable `Advanced` options
   * `Custom Model` to the prefix `litellm_proxy/` + the model you will be using (e.g. `litellm_proxy/anthropic.claude-3-5-sonnet-20241022-v2:0`)
   * `Base URL` to your LiteLLM proxy URL (e.g. `https://your-litellm-proxy.com`)
   * `API Key` to your LiteLLM proxy API key

docs/modules/usage/llms/llms.md

@@ -38,7 +38,7 @@ The following can be set in the OpenHands UI through the Settings:
 - `LLM Provider`
 - `LLM Model`
 - `API Key`
-- `Base URL` (through `Advanced Settings`)
+- `Base URL` (through `Advanced` settings)
 
 There are some settings that may be necessary for some LLMs/providers that cannot be set through the UI. Instead, these
 can be set through environment variables passed to the [docker run command](/modules/usage/installation#start-the-app)
@@ -63,22 +63,22 @@ We have a few guides for running OpenHands with specific model providers:
 ### API retries and rate limits
 
 LLM providers typically have rate limits, sometimes very low, and may require retries. OpenHands will automatically
-retry requests if it receives a Rate Limit Error (429 error code), API connection error, or other transient errors.
+retry requests if it receives a Rate Limit Error (429 error code).
 
 You can customize these options as you need for the provider you're using. Check their documentation, and set the
 following environment variables to control the number of retries and the time between retries:
 
-- `LLM_NUM_RETRIES` (Default of 8)
-- `LLM_RETRY_MIN_WAIT` (Default of 15 seconds)
-- `LLM_RETRY_MAX_WAIT` (Default of 120 seconds)
+- `LLM_NUM_RETRIES` (Default of 4 times)
+- `LLM_RETRY_MIN_WAIT` (Default of 5 seconds)
+- `LLM_RETRY_MAX_WAIT` (Default of 30 seconds)
 - `LLM_RETRY_MULTIPLIER` (Default of 2)
 
 If you are running OpenHands in development mode, you can also set these options in the `config.toml` file:
 
 ```toml
 [llm]
-num_retries = 8
-retry_min_wait = 15
-retry_max_wait = 120
+num_retries = 4
+retry_min_wait = 5
+retry_max_wait = 30
 retry_multiplier = 2
 ```

docs/modules/usage/llms/openai-llms.md

@@ -8,7 +8,7 @@ When running OpenHands, you'll need to set the following in the OpenHands UI thr
 * `LLM Provider` to `OpenAI`
 * `LLM Model` to the model you will be using.
 [Visit here to see a full list of OpenAI models that LiteLLM supports.](https://docs.litellm.ai/docs/providers/openai#openai-chat-completion-models)
-If the model is not in the list, toggle `Advanced Options`, and enter it in `Custom Model` (e.g. openai/<model-name> like `openai/gpt-4o`).
+If the model is not in the list, toggle `Advanced` options, and enter it in `Custom Model` (e.g. openai/<model-name> like `openai/gpt-4o`).
 * `API Key` to your OpenAI API key. To find or create your OpenAI Project API Key, [see here](https://platform.openai.com/api-keys).
 
 ## Using OpenAI-Compatible Endpoints
@@ -18,7 +18,7 @@ Just as for OpenAI Chat completions, we use LiteLLM for OpenAI-compatible endpoi
 ## Using an OpenAI Proxy
 
 If you're using an OpenAI proxy, in the OpenHands UI through the Settings:
-1. Enable `Advanced Options`
+1. Enable `Advanced` options
 2. Set the following:
    - `Custom Model` to openai/<model-name> (e.g. `openai/gpt-4o` or openai/<proxy-prefix>/<model-name>)
    - `Base URL` to the URL of your OpenAI proxy

docs/modules/usage/llms/openrouter.md

@@ -8,5 +8,5 @@ When running OpenHands, you'll need to set the following in the OpenHands UI thr
 * `LLM Provider` to `OpenRouter`
 * `LLM Model` to the model you will be using.
 [Visit here to see a full list of OpenRouter models](https://openrouter.ai/models).
-If the model is not in the list, toggle `Advanced Options`, and enter it in `Custom Model` (e.g. openrouter/<model-name> like `openrouter/anthropic/claude-3.5-sonnet`).
+If the model is not in the list, toggle `Advanced` options, and enter it in `Custom Model` (e.g. openrouter/<model-name> like `openrouter/anthropic/claude-3.5-sonnet`).
 * `API Key` to your OpenRouter API key.

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:0.21-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/sidebars.ts

@@ -42,7 +42,7 @@ const sidebars: SidebarsConfig = {
               id: 'usage/prompting/microagents-public',
             },
           ],
-        }
+        },
       ],
     },
     {
@@ -66,9 +66,26 @@ const sidebars: SidebarsConfig = {
         },
         {
           type: 'doc',
-          label: 'Github Actions',
+          label: 'Github Action',
           id: 'usage/how-to/github-action',
         },
+        {
+          type: 'category',
+          label: 'Cloud',
+          items: [
+            {
+              type: 'doc',
+              label: 'Openhands Cloud',
+              id: 'usage/cloud/openhands-cloud',
+            },
+
+            {
+              type: 'doc',
+              label: 'Cloud GitHub Resolver',
+              id: 'usage/cloud/cloud-github-resolver',
+            },
+          ],
+        },
       ],
     },
     {
@@ -185,7 +202,7 @@ const sidebars: SidebarsConfig = {
       type: 'doc',
       label: 'About',
       id: 'usage/about',
-    }
+    },
   ],
 };
 

docs/src/pages/index.tsx

@@ -23,6 +23,17 @@ export default function Home(): JSX.Element {
       })}
     >
       <HomepageHeader />
+      <div style={{ textAlign: 'center', padding: '2rem' }}>
+        <br />
+        <h2>Most Popular Links</h2>
+        <ul style={{ listStyleType: 'none'}}>
+          <li><a href="/modules/usage/Installation">How to Run OpenHands</a></li>
+          <li><a href="/modules/usage/prompting/microagents-repo">Customizing OpenHands to a repository</a></li>
+          <li><a href="/modules/usage/how-to/github-action">Integrating OpenHands with Github</a></li>
+          <li><a href="/modules/usage/llms#model-recommendations">Recommended models to use</a></li>
+          <li><a href="/modules/usage/runtimes#connecting-to-your-filesystem">Connecting OpenHands to your filesystem</a></li>
+        </ul>
+      </div>
     </Layout>
   );
 }

evaluation/README.md

@@ -20,6 +20,8 @@ To evaluate an agent, you can provide the agent's name to the `run_infer.py` pro
 ### Evaluating Different LLMs
 
 OpenHands in development mode uses `config.toml` to keep track of most configuration.
+**IMPORTANT: For evaluation, only the LLM section in `config.toml` will be used. Other configurations, such as `save_trajectory_path`, are not applied during evaluation.**
+
 Here's an example configuration file you can use to define and use multiple LLMs:
 
 ```toml
@@ -40,6 +42,8 @@ api_key = "XXX"
 temperature = 0.0
 ```
 
+For other configurations specific to evaluation, such as `save_trajectory_path`, these are typically set in the `get_config` function of the respective `run_infer.py` file for each benchmark.
+
 ## Supported Benchmarks
 
 The OpenHands evaluation harness supports a wide variety of benchmarks across [software engineering](#software-engineering), [web browsing](#web-browsing), [miscellaneous assistance](#misc-assistance), and [real-world](#real-world) tasks.

evaluation/benchmarks/EDA/run_infer.py

@@ -9,6 +9,7 @@ from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -17,7 +18,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -60,16 +60,14 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=False,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -17,6 +17,7 @@ from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -25,7 +26,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     parse_arguments,
 )
@@ -40,20 +40,15 @@ from openhands.utils.async_utils import call_async_from_sync
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-slim'
+
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime=os.environ.get('RUNTIME', 'docker'),
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-slim',
-            enable_auto_lint=True,
-            use_host_network=False,
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-            remote_runtime_api_url=os.environ.get('SANDBOX_REMOTE_RUNTIME_API_URL'),
-            keep_runtime_alive=False,
-            remote_runtime_init_timeout=3600,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -16,6 +16,7 @@ from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -24,7 +25,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     load_from_toml,
     parse_arguments,
@@ -47,21 +47,14 @@ SKIP_NUM = (
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.11-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime=os.environ.get('RUNTIME', 'docker'),
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.11-bookworm',
-            enable_auto_lint=True,
-            use_host_network=False,
-            timeout=100,
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-            remote_runtime_api_url=os.environ.get('SANDBOX_REMOTE_RUNTIME_API_URL'),
-            keep_runtime_alive=False,
-            remote_runtime_init_timeout=1800,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -14,6 +14,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -22,7 +23,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     parse_arguments,
 )
@@ -57,17 +57,15 @@ def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
     BIOCODER_BENCH_CONTAINER_IMAGE = 'public.ecr.aws/i5g0m1f6/eval_biocoder:v1.0'
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = BIOCODER_BENCH_CONTAINER_IMAGE
 
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image=BIOCODER_BENCH_CONTAINER_IMAGE,
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -17,6 +17,7 @@ from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -25,7 +26,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     parse_arguments,
 )
@@ -71,16 +71,15 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
+
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -10,6 +10,7 @@ from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -18,7 +19,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     parse_arguments,
 )
@@ -36,16 +36,14 @@ def get_config(
     assert (
         metadata.max_iterations == 1
     ), 'max_iterations must be 1 for browsing delegation evaluation.'
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=False,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         workspace_base=None,
         workspace_mount_path=None,
     )

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/run_infer.py

@@ -15,6 +15,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     assert_and_raise,
     codeact_user_response,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -25,7 +26,6 @@ from openhands.controller.state.state import State
 from openhands.core.config import (
     AgentConfig,
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -105,9 +105,7 @@ def get_config(
     instance: pd.Series,
     metadata: EvalMetadata,
 ) -> AppConfig:
-    # COMMIT0_CONTAINER_IMAGE = 'wentingzhao/'
     assert USE_INSTANCE_IMAGE
-    # We use a different instance image for the each instance of commit0 eval
     repo_name = instance['repo'].split('/')[1]
     base_container_image = get_instance_docker_image(repo_name)
     logger.info(
@@ -115,27 +113,16 @@ def get_config(
         f'Please make sure this image exists. '
         f'Submit an issue on https://github.com/All-Hands-AI/OpenHands if you run into any issues.'
     )
-    # else:
-    #     raise
-    # base_container_image = SWE_BENCH_CONTAINER_IMAGE
-    # logger.info(f'Using swe-bench container image: {base_container_image}')
+
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = base_container_image
 
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         max_iterations=metadata.max_iterations,
         runtime=os.environ.get('RUNTIME', 'docker'),
-        sandbox=SandboxConfig(
-            base_container_image=base_container_image,
-            enable_auto_lint=True,
-            use_host_network=False,
-            # large enough timeout, since some testcases take very long to run
-            timeout=300,
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-            remote_runtime_api_url=os.environ.get('SANDBOX_REMOTE_RUNTIME_API_URL'),
-            keep_runtime_alive=False,
-            remote_runtime_init_timeout=3600,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -16,6 +16,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -25,7 +26,6 @@ from openhands.controller.state.state import State
 from openhands.core.config import (
     AgentConfig,
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     parse_arguments,
 )
@@ -62,16 +62,14 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -13,6 +13,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -21,10 +22,10 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
+from openhands.core.config.utils import get_agent_config_arg
 from openhands.core.logger import openhands_logger as logger
 from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import AgentFinishAction, CmdRunAction, MessageAction
@@ -47,23 +48,25 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,
     )
     config.set_llm_config(metadata.llm_config)
-    agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.enable_prompt_extensions = False
+    if metadata.agent_config:
+        config.set_agent_config(metadata.agent_config, metadata.agent_class)
+    else:
+        logger.info('Agent config not provided, using default settings')
+        agent_config = config.get_agent_config(metadata.agent_class)
+        agent_config.enable_prompt_extensions = False
     return config
 
 
@@ -237,6 +240,10 @@ if __name__ == '__main__':
     )
     args, _ = parser.parse_known_args()
 
+    agent_config = None
+    if args.agent_config:
+        agent_config = get_agent_config_arg(args.agent_config)
+
     llm_config = None
     if args.llm_config:
         llm_config = get_llm_config_arg(args.llm_config)
@@ -255,6 +262,7 @@ if __name__ == '__main__':
         eval_output_dir=args.eval_output_dir,
         data_split=args.data_split,
         details={'gaia-level': args.level},
+        agent_config=agent_config,
     )
 
     dataset = load_dataset('gaia-benchmark/GAIA', args.level)

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"
@@ -9,6 +9,7 @@ AGENT=$3
 EVAL_LIMIT=$4
 LEVELS=$5
 NUM_WORKERS=$6
+AGENT_CONFIG=$7
 
 if [ -z "$NUM_WORKERS" ]; then
   NUM_WORKERS=1
@@ -49,5 +50,9 @@ if [ -n "$EVAL_LIMIT" ]; then
   COMMAND="$COMMAND --eval-n-limit $EVAL_LIMIT"
 fi
 
+if [ -n "$AGENT_CONFIG" ]; then
+  echo "AGENT_CONFIG: $AGENT_CONFIG"
+  COMMAND="$COMMAND --agent-config $AGENT_CONFIG"
+
 # Run the command
 eval $COMMAND

evaluation/benchmarks/gorilla/run_infer.py

@@ -11,6 +11,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -19,7 +20,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -40,16 +40,14 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -29,6 +29,7 @@ from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -37,7 +38,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -61,16 +61,14 @@ ACTION_FORMAT = """
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -22,6 +22,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -30,7 +31,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     parse_arguments,
 )
@@ -82,16 +82,14 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'python:3.12-bookworm'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='python:3.12-bookworm',
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -9,6 +9,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -17,7 +18,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -45,17 +45,18 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'xingyaoww/od-eval-logic-reasoning:v1.0'
+    sandbox_config.runtime_extra_deps = (
+        '$OH_INTERPRETER_PATH -m pip install scitools-pyke'
+    )
+
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='xingyaoww/od-eval-logic-reasoning:v1.0',
-            enable_auto_lint=True,
-            use_host_network=False,
-            runtime_extra_deps='$OH_INTERPRETER_PATH -m pip install scitools-pyke',
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -12,6 +12,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -21,7 +22,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     parse_arguments,
 )
@@ -55,22 +55,14 @@ def get_config(
     metadata: EvalMetadata,
     env_id: str,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'xingyaoww/od-eval-miniwob:v1.0'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime=os.environ.get('RUNTIME', 'docker'),
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='xingyaoww/od-eval-miniwob:v1.0',
-            enable_auto_lint=True,
-            use_host_network=False,
-            browsergym_eval_env=env_id,
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-            remote_runtime_api_url=os.environ.get('SANDBOX_REMOTE_RUNTIME_API_URL'),
-            remote_runtime_init_timeout=1800,
-            keep_runtime_alive=False,
-            timeout=120,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -14,6 +14,7 @@ from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -22,7 +23,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -103,17 +103,18 @@ def load_incontext_example(task_name: str, with_tool: bool = True):
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'xingyaoww/od-eval-mint:v1.0'
+    sandbox_config.runtime_extra_deps = (
+        f'$OH_INTERPRETER_PATH -m pip install {" ".join(MINT_DEPENDENCIES)}'
+    )
+
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='xingyaoww/od-eval-mint:v1.0',
-            enable_auto_lint=True,
-            use_host_network=False,
-            runtime_extra_deps=f'$OH_INTERPRETER_PATH -m pip install {" ".join(MINT_DEPENDENCIES)}',
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -25,6 +25,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -33,7 +34,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
     load_app_config,
@@ -77,16 +77,14 @@ ID2CONDA = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = 'public.ecr.aws/i5g0m1f6/ml-bench'
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime='docker',
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='public.ecr.aws/i5g0m1f6/ml-bench',
-            enable_auto_lint=True,
-            use_host_network=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/run_infer.py

@@ -11,6 +11,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     codeact_user_response,
     compatibility_for_eval_history_pairs,
+    get_default_sandbox_config_for_eval,
     make_metadata,
     prepare_dataset,
     reset_logger_for_multiprocessing,
@@ -20,7 +21,6 @@ from evaluation.utils.shared import (
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -59,21 +59,17 @@ def get_config(
     metadata: EvalMetadata,
     instance_id: str,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = (
+        'docker.io/xingyaoww/openhands-eval-scienceagentbench'
+    )
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         runtime=os.environ.get('RUNTIME', 'docker'),
         max_budget_per_task=4,
         max_iterations=metadata.max_iterations,
-        sandbox=SandboxConfig(
-            base_container_image='docker.io/xingyaoww/openhands-eval-scienceagentbench',
-            enable_auto_lint=True,
-            use_host_network=False,
-            timeout=300,
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-            remote_runtime_api_url=os.environ.get('SANDBOX_REMOTE_RUNTIME_API_URL'),
-            keep_runtime_alive=False,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,

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/eval_infer.py

@@ -1,5 +1,6 @@
 import json
 import os
+import subprocess
 import tempfile
 import time
 from functools import partial
@@ -21,13 +22,14 @@ from evaluation.benchmarks.swe_bench.run_infer import get_instance_docker_image
 from evaluation.utils.shared import (
     EvalMetadata,
     EvalOutput,
+    get_default_sandbox_config_for_eval,
     prepare_dataset,
     reset_logger_for_multiprocessing,
     run_evaluation,
 )
 from openhands.core.config import (
     AppConfig,
-    SandboxConfig,
+    LLMConfig,
     get_parser,
 )
 from openhands.core.logger import openhands_logger as logger
@@ -79,22 +81,16 @@ def get_config(metadata: EvalMetadata, instance: pd.Series) -> AppConfig:
         f'Please make sure this image exists. '
         f'Submit an issue on https://github.com/All-Hands-AI/OpenHands if you run into any issues.'
     )
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = base_container_image
+    sandbox_config.remote_runtime_resource_factor = get_instance_resource_factor(
+        dataset_name=metadata.dataset,
+        instance_id=instance['instance_id'],
+    )
     config = AppConfig(
         run_as_openhands=False,
         runtime=os.environ.get('RUNTIME', 'docker'),
-        sandbox=SandboxConfig(
-            base_container_image=base_container_image,
-            use_host_network=False,
-            # large enough timeout, since some testcases take very long to run
-            timeout=600,
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-            remote_runtime_api_url=os.environ.get('SANDBOX_REMOTE_RUNTIME_API_URL'),
-            remote_runtime_init_timeout=3600,
-            remote_runtime_resource_factor=get_instance_resource_factor(
-                dataset_name=metadata.dataset,
-                instance_id=instance['instance_id'],
-            ),
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,
@@ -412,6 +408,21 @@ if __name__ == '__main__':
         with open(metadata_filepath, 'r') as metadata_file:
             data = metadata_file.read()
             metadata = EvalMetadata.model_validate_json(data)
+    else:
+        # Initialize with a dummy metadata when file doesn't exist
+        metadata = EvalMetadata(
+            agent_class='dummy_agent',  # Placeholder agent class
+            llm_config=LLMConfig(model='dummy_model'),  # Minimal LLM config
+            max_iterations=1,  # Minimal iterations
+            eval_output_dir=os.path.dirname(
+                args.input_file
+            ),  # Use input file dir as output dir
+            start_time=time.strftime('%Y-%m-%d %H:%M:%S'),  # Current time
+            git_commit=subprocess.check_output(['git', 'rev-parse', 'HEAD'])
+            .decode('utf-8')
+            .strip(),  # Current commit
+            dataset=args.dataset,  # Dataset name from args
+        )
 
     # The evaluation harness constrains the signature of `process_instance_func` but we need to
     # pass extra information. Build a new function object to avoid issues with multiprocessing.

evaluation/benchmarks/swe_bench/resource/mapping.py

@@ -7,6 +7,7 @@ This file tracks the resource requirements of different instances.
 
 import json
 import os
+
 from openhands.core.logger import openhands_logger as logger
 
 CUR_DIR = os.path.dirname(os.path.abspath(__file__))

evaluation/benchmarks/swe_bench/resource/princeton-nlp__SWE-bench_Verified-test.json

@@ -1 +0,0 @@
-{"pydata__xarray-6721": 8, "pytest-dev__pytest-7236": 8, "matplotlib__matplotlib-24627": 4, "django__django-15561": 4, "django__django-15098": 4, "django__django-14771": 4, "sympy__sympy-21612": 4, "sympy__sympy-15345": 4, "psf__requests-5414": 4, "astropy__astropy-14508": 2, "django__django-11451": 2, "django__django-11477": 2, "django__django-10880": 2, "django__django-11163": 2, "django__django-11815": 2, "astropy__astropy-14369": 2, "django__django-10097": 2, "django__django-10554": 2, "django__django-12304": 2, "django__django-12325": 2, "django__django-11551": 2, "django__django-11734": 2, "django__django-13109": 2, "django__django-13089": 2, "django__django-13343": 2, "django__django-13363": 2, "django__django-13809": 2, "django__django-13810": 2, "django__django-13786": 2, "django__django-13807": 2, "django__django-14493": 2, "django__django-11820": 2, "django__django-11951": 2, "django__django-11964": 2, "astropy__astropy-14309": 2, "astropy__astropy-14365": 2, "astropy__astropy-12907": 2, "astropy__astropy-14182": 2, "django__django-15161": 2, "django__django-15128": 2, "django__django-14999": 2, "django__django-14915": 2, "django__django-14752": 2, "django__django-14765": 2, "django__django-14089": 2, "django__django-15252": 2, "django__django-15380": 2, "django__django-15382": 2, "django__django-15499": 2, "django__django-15467": 2, "django__django-15280": 2, "django__django-15315": 2, "django__django-15277": 2, "django__django-15268": 2, "django__django-15629": 2, "django__django-15695": 2, "django__django-15732": 2, "django__django-15863": 2, "django__django-16082": 2, "django__django-16145": 2, "django__django-16256": 2, "django__django-16429": 2, "django__django-16454": 2, "django__django-16493": 2, "matplotlib__matplotlib-13989": 2, "matplotlib__matplotlib-20488": 2, "django__django-15503": 2, "django__django-15525": 2, "django__django-15375": 2, "django__django-15278": 2, "matplotlib__matplotlib-21568": 2, "matplotlib__matplotlib-20859": 2, "matplotlib__matplotlib-20826": 2, "matplotlib__matplotlib-20676": 2, "matplotlib__matplotlib-23412": 2, "matplotlib__matplotlib-22719": 2, "matplotlib__matplotlib-23299": 2, "matplotlib__matplotlib-22865": 2, "matplotlib__matplotlib-24149": 2, "matplotlib__matplotlib-24177": 2, "matplotlib__matplotlib-24570": 2, "matplotlib__matplotlib-24637": 2, "matplotlib__matplotlib-24970": 2, "matplotlib__matplotlib-23476": 2, "matplotlib__matplotlib-24026": 2, "matplotlib__matplotlib-23314": 2, "matplotlib__matplotlib-25332": 2, "matplotlib__matplotlib-25311": 2, "matplotlib__matplotlib-25122": 2, "matplotlib__matplotlib-25479": 2, "matplotlib__matplotlib-26342": 2, "psf__requests-2317": 2, "matplotlib__matplotlib-25960": 2, "matplotlib__matplotlib-25775": 2, "pydata__xarray-4356": 2, "pydata__xarray-4075": 2, "pydata__xarray-6461": 2, "pydata__xarray-4687": 2, "pydata__xarray-6599": 2, "pylint-dev__pylint-4661": 2, "django__django-15554": 2, "django__django-15563": 2, "pytest-dev__pytest-5262": 2, "pytest-dev__pytest-10081": 2, "scikit-learn__scikit-learn-12973": 2, "scikit-learn__scikit-learn-13124": 2, "scikit-learn__scikit-learn-13779": 2, "scikit-learn__scikit-learn-14141": 2, "scikit-learn__scikit-learn-13439": 2, "scikit-learn__scikit-learn-13496": 2, "scikit-learn__scikit-learn-15100": 2, "scikit-learn__scikit-learn-25102": 2, "scikit-learn__scikit-learn-25232": 2, "scikit-learn__scikit-learn-25747": 2, "scikit-learn__scikit-learn-26323": 2, "scikit-learn__scikit-learn-9288": 2, "scikit-learn__scikit-learn-14496": 2, "scikit-learn__scikit-learn-14629": 2, "sphinx-doc__sphinx-8265": 2, "sphinx-doc__sphinx-8548": 2, "sphinx-doc__sphinx-8593": 2, "sphinx-doc__sphinx-8595": 2, "sphinx-doc__sphinx-8621": 2, "sphinx-doc__sphinx-8638": 2, "sphinx-doc__sphinx-9229": 2, "sphinx-doc__sphinx-9281": 2, "sphinx-doc__sphinx-9461": 2, "sphinx-doc__sphinx-9591": 2, "sphinx-doc__sphinx-9658": 2, "sphinx-doc__sphinx-9673": 2, "sympy__sympy-12096": 2, "sympy__sympy-12481": 2, "sphinx-doc__sphinx-10323": 2, "sphinx-doc__sphinx-7590": 2, "sympy__sympy-13877": 2, "sympy__sympy-12489": 2, "sympy__sympy-15809": 2, "sympy__sympy-14711": 2, "sympy__sympy-16597": 2, "sympy__sympy-16766": 2, "sympy__sympy-16792": 2, "sympy__sympy-15875": 2, "sympy__sympy-17655": 2, "sympy__sympy-18189": 2, "sympy__sympy-18763": 2, "sympy__sympy-19040": 2, "sympy__sympy-19495": 2, "sympy__sympy-19637": 2, "sympy__sympy-19783": 2, "sympy__sympy-17630": 2, "sympy__sympy-20428": 2, "sympy__sympy-20590": 2, "sympy__sympy-20801": 2, "sympy__sympy-21379": 2, "sympy__sympy-21847": 2, "sympy__sympy-22456": 2, "sympy__sympy-22714": 2, "sympy__sympy-22914": 2, "sympy__sympy-23262": 2, "sympy__sympy-23413": 2, "sympy__sympy-23534": 2, "sympy__sympy-24066": 2, "sympy__sympy-24213": 2, "sympy__sympy-24443": 2, "sympy__sympy-24562": 2, "sympy__sympy-24661": 2}

evaluation/benchmarks/swe_bench/run_infer.py

@@ -18,6 +18,7 @@ from evaluation.utils.shared import (
     EvalOutput,
     assert_and_raise,
     codeact_user_response,
+    get_default_sandbox_config_for_eval,
     get_metrics,
     is_fatal_evaluation_error,
     make_metadata,
@@ -30,7 +31,6 @@ from openhands.controller.state.state import State
 from openhands.core.config import (
     AgentConfig,
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -122,28 +122,23 @@ def get_config(
         base_container_image = SWE_BENCH_CONTAINER_IMAGE
         logger.info(f'Using swe-bench container image: {base_container_image}')
 
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = base_container_image
+    sandbox_config.enable_auto_lint = True
+    sandbox_config.use_host_network = False
+    # Add platform to the sandbox config to solve issue 4401
+    sandbox_config.platform = 'linux/amd64'
+    sandbox_config.remote_runtime_resource_factor = get_instance_resource_factor(
+        dataset_name=metadata.dataset,
+        instance_id=instance['instance_id'],
+    )
+
     config = AppConfig(
         default_agent=metadata.agent_class,
         run_as_openhands=False,
         max_iterations=metadata.max_iterations,
         runtime=os.environ.get('RUNTIME', 'docker'),
-        sandbox=SandboxConfig(
-            base_container_image=base_container_image,
-            enable_auto_lint=True,
-            use_host_network=False,
-            # large enough timeout, since some testcases take very long to run
-            timeout=300,
-            # Add platform to the sandbox config to solve issue 4401
-            platform='linux/amd64',
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-            remote_runtime_api_url=os.environ.get('SANDBOX_REMOTE_RUNTIME_API_URL'),
-            keep_runtime_alive=False,
-            remote_runtime_init_timeout=3600,
-            remote_runtime_resource_factor=get_instance_resource_factor(
-                dataset_name=metadata.dataset,
-                instance_id=instance['instance_id'],
-            ),
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,
@@ -329,6 +324,22 @@ def complete_runtime(
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
+
+    if obs.exit_code == -1:
+        # The previous command is still running
+        # We need to kill previous command
+        logger.info('The previous command is still running, trying to kill it...')
+        action = CmdRunAction(command='C-c')
+        obs = runtime.run_action(action)
+        logger.info(obs, extra={'msg_type': 'OBSERVATION'})
+
+        # Then run the command again
+        action = CmdRunAction(command=f'cd /workspace/{workspace_dir_name}')
+        action.set_hard_timeout(600)
+        logger.info(action, extra={'msg_type': 'ACTION'})
+        obs = runtime.run_action(action)
+        logger.info(obs, extra={'msg_type': 'OBSERVATION'})
+
     assert_and_raise(
         isinstance(obs, CmdOutputObservation) and obs.exit_code == 0,
         f'Failed to cd to /workspace/{workspace_dir_name}: {str(obs)}',

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