Save trajectory to workspace/trajectory.json after agent run

Co-authored-by: Engel Nyst <enyst@users.noreply.github.com>

.github/ISSUE_TEMPLATE/bug_template.yml

@@ -30,6 +30,7 @@ body:
       description: How are you running OpenHands?
       options:
         - Docker command in README
+        - GitHub resolver
         - Development workflow
         - app.all-hands.dev
         - Other

.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/integration-runner.yml

@@ -160,7 +160,6 @@ jobs:
           echo "api_key = \"$LLM_API_KEY\"" >> config.toml
           echo "base_url = \"$LLM_BASE_URL\"" >> config.toml
           echo "temperature = 0.0" >> config.toml
-
       - name: Run integration test evaluation for DelegatorAgent (DeepSeek)
         env:
           SANDBOX_FORCE_REBUILD_RUNTIME: True
@@ -174,12 +173,42 @@ jobs:
           cat $REPORT_FILE_DELEGATOR_DEEPSEEK >> $GITHUB_ENV
           echo >> $GITHUB_ENV
           echo "EOF" >> $GITHUB_ENV
+      # -------------------------------------------------------------
+      # Run VisualBrowsingAgent tests for DeepSeek, limited to t05 and t06
+      - name: Wait a little bit (again)
+        run: sleep 5
+
+      - name: Configure config.toml for testing VisualBrowsingAgent (DeepSeek)
+        env:
+          LLM_MODEL: "litellm_proxy/deepseek-chat"
+          LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
+          LLM_BASE_URL: ${{ secrets.LLM_BASE_URL }}
+          MAX_ITERATIONS: 15
+        run: |
+          echo "[llm.eval]" > config.toml
+          echo "model = \"$LLM_MODEL\"" >> config.toml
+          echo "api_key = \"$LLM_API_KEY\"" >> config.toml
+          echo "base_url = \"$LLM_BASE_URL\"" >> config.toml
+          echo "temperature = 0.0" >> config.toml
+      - name: Run integration test evaluation for VisualBrowsingAgent (DeepSeek)
+        env:
+          SANDBOX_FORCE_REBUILD_RUNTIME: True
+        run: |
+          poetry run ./evaluation/integration_tests/scripts/run_infer.sh llm.eval HEAD VisualBrowsingAgent '' 15 $N_PROCESSES "t05_simple_browsing,t06_github_pr_browsing.py" 'visualbrowsing_deepseek_run'
+
+          # Find and export the visual browsing agent test results
+          REPORT_FILE_VISUALBROWSING_DEEPSEEK=$(find evaluation/evaluation_outputs/outputs/integration_tests/VisualBrowsingAgent/deepseek*_maxiter_15_N* -name "report.md" -type f | head -n 1)
+          echo "REPORT_FILE_VISUALBROWSING_DEEPSEEK: $REPORT_FILE_VISUALBROWSING_DEEPSEEK"
+          echo "INTEGRATION_TEST_REPORT_VISUALBROWSING_DEEPSEEK<<EOF" >> $GITHUB_ENV
+          cat $REPORT_FILE_VISUALBROWSING_DEEPSEEK >> $GITHUB_ENV
+          echo >> $GITHUB_ENV
+          echo "EOF" >> $GITHUB_ENV
 
       - name: Create archive of evaluation outputs
         run: |
           TIMESTAMP=$(date +'%y-%m-%d-%H-%M')
           cd evaluation/evaluation_outputs/outputs  # Change to the outputs directory
-          tar -czvf ../../../integration_tests_${TIMESTAMP}.tar.gz integration_tests/CodeActAgent/* integration_tests/DelegatorAgent/*  # Only include the actual result directories
+          tar -czvf ../../../integration_tests_${TIMESTAMP}.tar.gz integration_tests/CodeActAgent/* integration_tests/DelegatorAgent/* integration_tests/VisualBrowsingAgent/* # Only include the actual result directories
 
       - name: Upload evaluation results as artifact
         uses: actions/upload-artifact@v4
@@ -227,4 +256,7 @@ jobs:
               **Integration Tests Report Delegator (DeepSeek)**
               ${{ env.INTEGRATION_TEST_REPORT_DELEGATOR_DEEPSEEK }}
               ---
+              **Integration Tests Report VisualBrowsing (DeepSeek)**
+              ${{ env.INTEGRATION_TEST_REPORT_VISUALBROWSING_DEEPSEEK }}
+              ---
               Download testing outputs (includes both Haiku and DeepSeek results): [Download](${{ steps.upload_results_artifact.outputs.artifact-url }})

.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
@@ -84,8 +88,10 @@ jobs:
         run: |
           python -m pip index versions openhands-ai > openhands_versions.txt
           OPENHANDS_VERSION=$(head -n 1 openhands_versions.txt | awk '{print $2}' | tr -d '()')
-          echo "openhands-ai==${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: |
@@ -103,15 +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: ${{ github.token }}
@@ -168,7 +175,7 @@ jobs:
           echo "SANDBOX_ENV_BASE_CONTAINER_IMAGE=${{ 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
@@ -184,6 +191,7 @@ jobs:
             });
 
       - name: Install OpenHands
+        id: install_openhands
         uses: actions/github-script@v7
         env:
           COMMENT_BODY: ${{ github.event.comment.body || '' }}
@@ -196,7 +204,6 @@ jobs:
             const reviewBody = process.env.REVIEW_BODY.trim();
             const labelName = process.env.LABEL_NAME.trim();
             const eventName = process.env.EVENT_NAME.trim();
-
             // Check conditions
             const isExperimentalLabel = labelName === "fix-me-experimental";
             const isIssueCommentExperimental =
@@ -205,6 +212,9 @@ jobs:
             const isReviewCommentExperimental =
               eventName === "pull_request_review" && reviewBody.includes("@openhands-agent-exp");
 
+            // Set output variable
+            core.setOutput('isExperimental', isExperimentalLabel || isIssueCommentExperimental || isReviewCommentExperimental);
+
             // Perform package installation
             if (isExperimentalLabel || isIssueCommentExperimental || isReviewCommentExperimental) {
               console.log("Installing experimental OpenHands...");
@@ -213,16 +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 \
@@ -230,7 +242,8 @@ jobs:
             --issue-number ${{ env.ISSUE_NUMBER }} \
             --issue-type ${{ env.ISSUE_TYPE }} \
             --max-iterations ${{ env.MAX_ITERATIONS }} \
-            --comment-id ${{ env.COMMENT_ID }}
+            --comment-id ${{ env.COMMENT_ID }} \
+            --is-experimental ${{ steps.install_openhands.outputs.isExperimental }}
 
       - name: Check resolution result
         id: check_result
@@ -254,14 +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

.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

.gitignore

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

.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.

CODE_OF_CONDUCT.md

@@ -113,6 +113,20 @@ individual, or aggression toward or disparagement of classes of individuals.
 **Consequence**: A permanent ban from any sort of public interaction within the
 community.
 
+### Slack and Discord Etiquettes
+
+These Slack and Discord etiquette guidelines are designed to foster an inclusive, respectful, and productive environment for all community members. By following these best practices, we ensure effective communication and collaboration while minimizing disruptions. Let’s work together to build a supportive and welcoming community!
+
+- Communicate respectfully and professionally, avoiding sarcasm or harsh language, and remember that tone can be difficult to interpret in text.
+- Use threads for specific discussions to keep channels organized and easier to follow.
+- Tag others only when their input is critical or urgent, and use @here, @channel or @everyone sparingly to minimize disruptions.
+- Be patient, as open-source contributors and maintainers often have other commitments and may need time to respond.
+- Post questions or discussions in the most relevant channel (e.g., for [slack - #general](https://app.slack.com/client/T06P212QSEA/C06P5NCGSFP) for general topics, [slack - #questions](https://openhands-ai.slack.com/archives/C06U8UTKSAD) for queries/questions, [discord - #general](https://discord.com/channels/1222935860639563850/1222935861386018885)).
+- When asking for help or raising issues, include necessary details like links, screenshots, or clear explanations to provide context.
+- Keep discussions in public channels whenever possible to allow others to benefit from the conversation, unless the matter is sensitive or private.
+- Always adhere to [our standards](https://github.com/All-Hands-AI/OpenHands/blob/main/CODE_OF_CONDUCT.md#our-standards) to ensure a welcoming and collaborative environment.
+- If you choose to mute a channel, consider setting up alerts for topics that still interest you to stay engaged. For Slack, Go to Settings → Notifications → My Keywords to add specific keywords that will notify you when mentioned. For example, if you're here for discussions about LLMs, mute the channel if it’s too busy, but set notifications to alert you only when “LLMs” appears in messages. Also for Discord, go to the channel notifications and choose the option that best describes your need.
+
 ## Attribution
 
 This Code of Conduct is adapted from the [Contributor Covenant][homepage],

Development.md

@@ -5,7 +5,7 @@ Otherwise, you can clone the OpenHands project directly.
 
 ## Start the Server for Development
 ### 1. Requirements
-* Linux, Mac OS, or [WSL on Windows](https://learn.microsoft.com/en-us/windows/wsl/install)  [Ubuntu <= 22.04]
+* Linux, Mac OS, or [WSL on Windows](https://learn.microsoft.com/en-us/windows/wsl/install)  [Ubuntu >= 22.04]
 * [Docker](https://docs.docker.com/engine/install/) (For those on MacOS, make sure to allow the default Docker socket to be used from advanced settings!)
 * [Python](https://www.python.org/downloads/) = 3.12
 * [NodeJS](https://nodejs.org/en/download/package-manager) >= 20.x
@@ -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:0.20-nikolaik`
+Example: `export SANDBOX_RUNTIME_CONTAINER_IMAGE=ghcr.io/all-hands-ai/runtime:0.25-nikolaik`
 
 ## Develop inside Docker container
 

Makefile

@@ -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

@@ -12,7 +12,7 @@
   <a href="https://codecov.io/github/All-Hands-AI/OpenHands?branch=main"><img alt="CodeCov" src="https://img.shields.io/codecov/c/github/All-Hands-AI/OpenHands?style=for-the-badge&color=blue"></a>
   <a href="https://github.com/All-Hands-AI/OpenHands/blob/main/LICENSE"><img src="https://img.shields.io/github/license/All-Hands-AI/OpenHands?style=for-the-badge&color=blue" alt="MIT License"></a>
   <br/>
-  <a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw"><img src="https://img.shields.io/badge/Slack-Join%20Us-red?logo=slack&logoColor=white&style=for-the-badge" alt="Join our Slack community"></a>
+  <a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2ypg5jweb-d~6hObZDbXi_HEL8PDrbHg"><img src="https://img.shields.io/badge/Slack-Join%20Us-red?logo=slack&logoColor=white&style=for-the-badge" alt="Join our Slack community"></a>
   <a href="https://discord.gg/ESHStjSjD4"><img src="https://img.shields.io/badge/Discord-Join%20Us-purple?logo=discord&logoColor=white&style=for-the-badge" alt="Join our Discord community"></a>
   <a href="https://github.com/All-Hands-AI/OpenHands/blob/main/CREDITS.md"><img src="https://img.shields.io/badge/Project-Credits-blue?style=for-the-badge&color=FFE165&logo=github&logoColor=white" alt="Credits"></a>
   <br/>
@@ -39,21 +39,21 @@ Learn more at [docs.all-hands.dev](https://docs.all-hands.dev), or jump to the [
 ## ⚡ Quick Start
 
 The easiest way to run OpenHands is in Docker.
-See the [Installation](https://docs.all-hands.dev/modules/usage/installation) guide for
+See the [Running OpenHands](https://docs.all-hands.dev/modules/usage/installation) guide for
 system requirements and more information.
 
 ```bash
-docker pull docker.all-hands.dev/all-hands-ai/runtime:0.20-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20
+    docker.all-hands.dev/all-hands-ai/openhands:0.25
 ```
 
 You'll find OpenHands running at [http://localhost:3000](http://localhost:3000)!
@@ -69,7 +69,7 @@ run OpenHands in a scriptable [headless mode](https://docs.all-hands.dev/modules
 interact with it via a [friendly CLI](https://docs.all-hands.dev/modules/usage/how-to/cli-mode),
 or run it on tagged issues with [a github action](https://docs.all-hands.dev/modules/usage/how-to/github-action).
 
-Visit [Installation](https://docs.all-hands.dev/modules/usage/installation) for more information and setup instructions.
+Visit [Running OpenHands](https://docs.all-hands.dev/modules/usage/installation) for more information and setup instructions.
 
 > [!CAUTION]
 > OpenHands is meant to be run by a single user on their local workstation.
@@ -96,7 +96,7 @@ troubleshooting resources, and advanced configuration options.
 OpenHands is a community-driven project, and we welcome contributions from everyone. We do most of our communication
 through Slack, so this is the best place to start, but we also are happy to have you contact us on Discord or Github:
 
-- [Join our Slack workspace](https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw) - Here we talk about research, architecture, and future development.
+- [Join our Slack workspace](https://join.slack.com/t/openhands-ai/shared_invite/zt-2ypg5jweb-d~6hObZDbXi_HEL8PDrbHg) - Here we talk about research, architecture, and future development.
 - [Join our Discord server](https://discord.gg/ESHStjSjD4) - This is a community-run server for general discussion, questions, and feedback.
 - [Read or post Github Issues](https://github.com/All-Hands-AI/OpenHands/issues) - Check out the issues we're working on, or add your own ideas.
 

build.sh

@@ -1,5 +1,4 @@
 #!/bin/bash
 set -e
 
-cp pyproject.toml poetry.lock openhands
 poetry build -v

config.template.toml

@@ -23,6 +23,9 @@ workspace_base = "./workspace"
 # Cache directory path
 #cache_dir = "/tmp/cache"
 
+# Reasoning effort for o1 models (low, medium, high, or not set)
+#reasoning_effort = "medium"
+
 # Debugging enabled
 #debug = false
 
@@ -36,6 +39,11 @@ workspace_base = "./workspace"
 # If it's a folder, the session id will be used as the file name
 #save_trajectory_path="./trajectories"
 
+# Path to replay a trajectory, must be a file path
+# If provided, trajectory will be loaded and replayed before the
+# agent responds to any user instruction
+#replay_trajectory_path = ""
+
 # File store path
 #file_store_path = "/tmp/file_store"
 
@@ -67,7 +75,7 @@ workspace_base = "./workspace"
 #run_as_openhands = true
 
 # Runtime environment
-#runtime = "eventstream"
+#runtime = "docker"
 
 # Name of the default agent
 #default_agent = "CodeActAgent"
@@ -96,7 +104,7 @@ 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 = ""
@@ -187,7 +195,7 @@ model = "gpt-4o"
 #native_tool_calling = None
 
 [llm.gpt4o-mini]
-api_key = "your-api-key"
+api_key = ""
 model = "gpt-4o"
 
 
@@ -220,12 +228,16 @@ codeact_enable_jupyter = true
 # LLM config group to use
 #llm_config = 'your-llm-config-group'
 
-# Whether to use microagents at all
-#use_microagents = true
+# Whether to use prompt extension (e.g., microagent, repo/runtime info) at all
+#enable_prompt_extensions = 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/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:0.20-nikolaik}
+      - SANDBOX_RUNTIME_CONTAINER_IMAGE=${SANDBOX_RUNTIME_CONTAINER_IMAGE:-ghcr.io/all-hands-ai/runtime:0.25-nikolaik}
       - SANDBOX_USER_ID=${SANDBOX_USER_ID:-1234}
       - WORKSPACE_MOUNT_PATH=${WORKSPACE_BASE:-$PWD/workspace}
     ports:

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:0.20-nikolaik}
+      - SANDBOX_RUNTIME_CONTAINER_IMAGE=${SANDBOX_RUNTIME_CONTAINER_IMAGE:-docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik}
       #- SANDBOX_USER_ID=${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`
@@ -373,7 +373,7 @@ Les options de configuration de l'agent sont définies dans les sections `[agent
   - Description : Si l'éditeur LLM est activé dans l'espace d'action (fonctionne uniquement avec l'appel de fonction)
 
 **Utilisation du micro-agent**
-- `use_microagents`
+- `enable_prompt_extensions`
   - Type : `bool`
   - Valeur par défaut : `true`
   - Description : Indique si l'utilisation des micro-agents est activée ou non

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

@@ -95,7 +95,3 @@ sandbox_user_id="1001"
 ### Erreurs de port d'utilisation
 
 Si vous voyez un message d'erreur indiquant que le port est utilisé ou indisponible, essayez de supprimer toutes les containers docker en cours d'exécution (exécutez `docker ps` et `docker rm` des containers concernés) puis ré-exécutez ```make run```
-
-## Discuter
-
-Pour d'autres problèmes ou questions rejoignez le [Slack](https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw) ou le [Discord](https://discord.gg/ESHStjSjD4) et demandez!

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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
     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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
     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:0.20-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20
+    docker.all-hands.dev/all-hands-ai/openhands:0.25
 ```
 
 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/intro.mdx

@@ -42,7 +42,7 @@ Explorez le code source d'OpenHands sur [GitHub](https://github.com/All-Hands-AI
   />
 </a>
 <br></br>
-<a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw">
+<a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2ypg5jweb-d~6hObZDbXi_HEL8PDrbHg">
   <img
     src="https://img.shields.io/badge/Slack-Join%20Us-red?logo=slack&logoColor=white&style=for-the-badge"
     alt="Join our Slack community"

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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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`
@@ -373,7 +373,7 @@ Agent 配置选项在 `config.toml` 文件的 `[agent]` 和 `[agent.<agent_name>
   - 描述: 是否在 action space 中启用 Jupyter
 
 **Microagent 使用**
-- `use_microagents`
+- `enable_prompt_extensions`
   - 类型: `bool`
   - 默认值: `true`
   - 描述: 是否使用 microagents

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

@@ -96,7 +96,3 @@ sandbox_user_id="1001"
 ### 端口使用错误
 
 如果您遇到端口被占用或不可用的错误提示，可以尝试先用`docker ps`命令列出所有运行中的 Docker 容器，然后使用`docker rm`命令删除相关容器，最后再重新执行```make run```命令。
-
-## 讨论
-
-对于其他问题或疑问，请加入 [Slack](https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw) 或 [Discord](https://discord.gg/ESHStjSjD4) 提问！

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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
     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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
     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:0.20-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20
+    docker.all-hands.dev/all-hands-ai/openhands:0.25
 ```
 
 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/intro.mdx

@@ -42,7 +42,7 @@ OpenHands 是一个**自主 AI 软件工程师**，能够执行复杂的工程
   />
 </a>
 <br></br>
-<a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw">
+<a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2ypg5jweb-d~6hObZDbXi_HEL8PDrbHg">
   <img
     src="https://img.shields.io/badge/Slack-Join%20Us-red?logo=slack&logoColor=white&style=for-the-badge"
     alt="Join our Slack community"

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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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

@@ -55,6 +55,11 @@ The core configuration options are defined in the `[core]` section of the `confi
   - Default: `"./trajectories"`
   - Description: Path to store trajectories (can be a folder or a file). If it's a folder, the trajectories will be saved in a file named with the session id name and .json extension, in that folder.
 
+- `replay_trajectory_path`
+  - Type: `str`
+  - Default: `""`
+  - Description: Path to load a trajectory and replay. If given, must be a path to the trajectory file in JSON format. The actions in the trajectory file would be replayed first before any user instruction is executed.
+
 ### File Store
 - `file_store_path`
   - Type: `str`
@@ -121,7 +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`
@@ -335,8 +340,13 @@ 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
-- `use_microagents`
+- `enable_prompt_extensions`
   - Type: `bool`
   - Default: `true`
   - Description: Whether to use microagents at all

docs/modules/usage/getting-started.mdx

@@ -1,6 +1,6 @@
 # Getting Started with OpenHands
 
-So you've [installed OpenHands](./installation) and have
+So you've [run OpenHands](./installation) and have
 [set up your LLM](./installation#setup). Now what?
 
 OpenHands can help you tackle a wide variety of engineering tasks. But the technology

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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
     python -m openhands.core.cli
 ```
 

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

@@ -18,15 +18,21 @@ If you choose the first option, you can skip the `Create Your Docker Image` sect
 
 To create a custom Docker image, it must be Debian based.
 
-For example, if you want OpenHands to have `ruby` installed, create a `Dockerfile` with the following content:
+For example, if you want OpenHands to have `ruby` installed, you could create a `Dockerfile` with the following content:
 
 ```dockerfile
-FROM debian:latest
+FROM nikolaik/python-nodejs:python3.12-nodejs22
 
 # Install required packages
 RUN apt-get update && apt-get install -y ruby
 ```
 
+Or you could use a Ruby-specific base image:
+
+```dockerfile
+FROM ruby:latest
+```
+
 Save this file in a folder. Then, build your Docker image (e.g., named custom-image) by navigating to the folder in
 the terminal and running::
 ```bash
@@ -35,8 +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
 
@@ -55,6 +69,28 @@ This can be an image you’ve already pulled or one you’ve built:
 sandbox_base_container_image="custom-image"
 ```
 
+### Additional Configuration Options
+
+The `config.toml` file supports several other options for customizing your sandbox:
+
+```toml
+[core]
+# Install additional dependencies when the runtime is built
+# Can contain any valid shell commands
+# If you need the path to the Python interpreter in any of these commands, you can use the $OH_INTERPRETER_PATH variable
+runtime_extra_deps = """
+pip install numpy pandas
+apt-get update && apt-get install -y ffmpeg
+"""
+
+# Set environment variables for the runtime
+# Useful for configuration that needs to be available at runtime
+runtime_startup_env_vars = { DATABASE_URL = "postgresql://user:pass@localhost/db" }
+
+# Specify platform for multi-architecture builds (e.g., "linux/amd64" or "linux/arm64")
+platform = "linux/amd64"
+```
+
 ### Run
 
 Run OpenHands by running ```make run``` in the top level directory.

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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20 \
+    docker.all-hands.dev/all-hands-ai/openhands:0.25 \
     python -m openhands.core.main -t "write a bash script that prints hi"
 ```
 

docs/modules/usage/installation.mdx

@@ -1,27 +1,74 @@
-# Installation
+# Running OpenHands
 
 ## System Requirements
 
-- Docker version 26.0.0+ or Docker Desktop 4.31.0+.
-- You must be using Linux or Mac OS.
-  - If you are on Windows, you must use [WSL](https://learn.microsoft.com/en-us/windows/wsl/install).
+- MacOS with [Docker Desktop support](https://docs.docker.com/desktop/setup/install/mac-install/#system-requirements)
+- 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)
 
-## Start the app
+A system with a modern processor and a minimum of **4GB RAM** is recommended to run OpenHands.
+
+## Prerequisites
+
+<details>
+  <summary>MacOS</summary>
+
+  **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.
+</details>
+
+<details>
+  <summary>Linux</summary>
+
+  :::note
+  Tested with Ubuntu 22.04.
+  :::
+
+  **Docker Desktop**
+
+  1. [Install Docker Desktop on Linux](https://docs.docker.com/desktop/setup/install/linux/).
+
+</details>
+
+<details>
+  <summary>Windows</summary>
+
+  **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**
+
+  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
 
 The easiest way to run OpenHands is in Docker.
 
 ```bash
-docker pull docker.all-hands.dev/all-hands-ai/runtime:0.20-nikolaik
+docker pull docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik
 
 docker run -it --rm --pull=always \
-    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-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:0.20
+    docker.all-hands.dev/all-hands-ai/openhands:0.25
 ```
 
 You'll find OpenHands running at http://localhost:3000!
@@ -33,24 +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/custom-llm-configs.md

@@ -101,6 +101,36 @@ In this example:
 - Code generation uses GPT-4 with a higher token limit for generating larger code blocks
 - The default configuration remains available for other tasks
 
+# Custom Configurations with Reserved Names
+
+OpenHands can use custom LLM configurations named with reserved names, for specific use cases. If you specify the model and other settings under the reserved names, then OpenHands will load and them for a specific purpose. As of now, one such configuration is implemented: draft editor.
+
+## Draft Editor Configuration
+
+The `draft_editor` configuration is a group of settings you can provide, to specify the model to use for preliminary drafting of code edits, for any tasks that involve editing and refining code. You need to provide it under the section `[llm.draft_editor]`.
+
+For example, you can define in `config.toml` a draft editor like this:
+
+```toml
+[llm.draft_editor]
+model = "gpt-4"
+temperature = 0.2
+top_p = 0.95
+presence_penalty = 0.0
+frequency_penalty = 0.0
+```
+
+This configuration:
+- Uses GPT-4 for high-quality edits and suggestions
+- Sets a low temperature (0.2) to maintain consistency while allowing some flexibility
+- Uses a high top_p value (0.95) to consider a wide range of token options
+- Disables presence and frequency penalties to maintain focus on the specific edits needed
+
+Use this configuration when you want to let an LLM draft edits before making them. In general, it may be useful to:
+- Review and suggest code improvements
+- Refine existing content while maintaining its core meaning
+- Make precise, focused changes to code or text
+
 :::note
 Custom LLM configurations are only available when using OpenHands in development mode, via `main.py` or `cli.py`. When running via `docker run`, please use the standard configuration options.
 :::

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:0.20-nikolaik \
+    -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.25-nikolaik \
     -v /var/run/docker.sock:/var/run/docker.sock \
     # ...
 ```

docs/package.json

@@ -22,8 +22,8 @@
     "@mdx-js/react": "^3.1.0",
     "clsx": "^2.0.0",
     "prism-react-renderer": "^2.4.1",
-    "react": "^18.3.1",
-    "react-dom": "^18.3.1",
+    "react": "^19.0.0",
+    "react-dom": "^19.0.0",
     "react-icons": "^5.4.0",
     "react-use": "^17.6.0"
   },
@@ -31,7 +31,7 @@
     "@docusaurus/module-type-aliases": "^3.5.1",
     "@docusaurus/tsconfig": "^3.7.0",
     "@docusaurus/types": "^3.5.1",
-    "typescript": "~5.7.2"
+    "typescript": "~5.7.3"
   },
   "browserslist": {
     "production": [

docs/sidebars.ts

@@ -5,7 +5,7 @@ const sidebars: SidebarsConfig = {
   docsSidebar: [
     {
       type: 'doc',
-      label: 'Installation',
+      label: 'Running OpenHands',
       id: 'usage/installation',
     },
     {
@@ -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/components/CustomFooter.tsx

@@ -8,7 +8,7 @@ function CustomFooter() {
     <footer className="custom-footer">
       <div className="footer-content">
         <div className="footer-icons">
-          <a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw" target="_blank" rel="noopener noreferrer">
+          <a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2ypg5jweb-d~6hObZDbXi_HEL8PDrbHg" target="_blank" rel="noopener noreferrer">
             <FaSlack />
           </a>
           <a href="https://discord.gg/ESHStjSjD4" target="_blank" rel="noopener noreferrer">

docs/src/components/HomepageHeader/HomepageHeader.tsx

@@ -23,7 +23,7 @@ export function HomepageHeader() {
           <a href="https://codecov.io/github/All-Hands-AI/OpenHands?branch=main"><img alt="CodeCov" src="https://img.shields.io/codecov/c/github/All-Hands-AI/OpenHands?style=for-the-badge&color=blue" /></a>
           <a href="https://github.com/All-Hands-AI/OpenHands/blob/main/LICENSE"><img src="https://img.shields.io/github/license/All-Hands-AI/OpenHands?style=for-the-badge&color=blue" alt="MIT License" /></a>
           <br/>
-          <a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2wkh4pklz-w~h_DVDtEe9H5kyQlcNxVw"><img src="https://img.shields.io/badge/Slack-Join%20Us-red?logo=slack&logoColor=white&style=for-the-badge" alt="Join our Slack community" /></a>
+          <a href="https://join.slack.com/t/openhands-ai/shared_invite/zt-2ypg5jweb-d~6hObZDbXi_HEL8PDrbHg"><img src="https://img.shields.io/badge/Slack-Join%20Us-red?logo=slack&logoColor=white&style=for-the-badge" alt="Join our Slack community" /></a>
           <a href="https://discord.gg/ESHStjSjD4"><img src="https://img.shields.io/badge/Discord-Join%20Us-purple?logo=discord&logoColor=white&style=for-the-badge" alt="Join our Discord community" /></a>
           <a href="https://github.com/All-Hands-AI/OpenHands/blob/main/CREDITS.md"><img src="https://img.shields.io/badge/Project-Credits-blue?style=for-the-badge&color=FFE165&logo=github&logoColor=white" alt="Credits" /></a>
           <br/>

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,
 )
@@ -60,23 +60,21 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
 )
@@ -40,27 +40,22 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
@@ -47,28 +47,21 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
 
     # copy 'draft_editor' config if exists
     config_copy = copy.deepcopy(config)

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,
 )
@@ -57,24 +57,22 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
 )
@@ -71,23 +71,22 @@ 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.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
 )
@@ -36,22 +36,20 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
@@ -171,7 +158,7 @@ def initialize_runtime(
     action = CmdRunAction(
         command=f'git clone -b commit0_combined https://github.com/{instance["repo"]}.git'
     )
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -181,7 +168,7 @@ def initialize_runtime(
     )
 
     action = CmdRunAction(command=f'cd /workspace/{workspace_dir_name}')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -191,7 +178,7 @@ def initialize_runtime(
     )
 
     action = CmdRunAction(command='git checkout -b openhands')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -201,7 +188,7 @@ def initialize_runtime(
 
     # Install commit0
     action = CmdRunAction(command='/root/.cargo/bin/uv pip install commit0')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     # logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -231,7 +218,7 @@ def complete_runtime(
     workspace_dir_name = _get_commit0_workspace_dir_name(instance)
 
     action = CmdRunAction(command='git add .')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -241,7 +228,7 @@ def complete_runtime(
     )
 
     action = CmdRunAction(command='git commit -m "openhands edits"')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -258,7 +245,7 @@ def complete_runtime(
         action = CmdRunAction(
             command=f"git diff {instance['base_commit']} HEAD -- . ':(exclude)spec.pdf.bz2'"
         )
-        action.timeout = 600 + 100 * n_retries
+        action.set_hard_timeout(600 + 100 * n_retries)
         logger.info(action, extra={'msg_type': 'ACTION'})
         obs = runtime.run_action(action)
         # logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -282,7 +269,7 @@ def complete_runtime(
     action = CmdRunAction(
         command=f"{instance['test']['test_cmd']} --json-report --json-report-file=report.json --continue-on-collection-errors {test_dir} > test_output.txt 2>&1"
     )
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -292,7 +279,7 @@ def complete_runtime(
     )
     # Read test output
     action = CmdRunAction(command='cat test_output.txt')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     # logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -305,7 +292,7 @@ def complete_runtime(
 
     # Save pytest exit code
     action = CmdRunAction(command='echo $?')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     # logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -318,7 +305,7 @@ def complete_runtime(
 
     # Read the test report
     action = CmdRunAction(command='cat report.json')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     # logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -330,7 +317,7 @@ def complete_runtime(
     repo_name = instance['repo'].split('/')[1]
     repo_name = repo_name.replace('.', '-')
     action = CmdRunAction(command=f'commit0 get-tests {repo_name}')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     # logger.info(obs, extra={'msg_type': 'OBSERVATION'})

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,
 )
@@ -62,23 +62,21 @@ 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.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     agent_config = AgentConfig(
         function_calling=False,
         codeact_enable_jupyter=True,

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.use_microagents = 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

@@ -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,
 )
@@ -40,23 +40,21 @@ 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.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
 )
@@ -61,23 +61,21 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
 )
@@ -82,23 +82,21 @@ 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.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
 )
@@ -45,24 +45,25 @@ AGENT_CLS_TO_INST_SUFFIX = {
 def get_config(
     metadata: EvalMetadata,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = '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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

evaluation/benchmarks/miniwob/README.md

@@ -8,6 +8,9 @@ Please follow instruction [here](../../README.md#setup) to setup your local deve
 
 ## Test if your environment works
 
+Follow the instructions here https://miniwob.farama.org/content/getting_started/ & https://miniwob.farama.org/content/viewing/
+to set up MiniWoB server in your local environment at http://localhost:8080/miniwob/
+
 Access with browser the above MiniWoB URLs and see if they load correctly.
 
 ## Run Evaluation

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/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,
 )
@@ -103,24 +103,25 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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,
@@ -77,23 +77,21 @@ 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,
     )
     config.set_llm_config(metadata.llm_config)
     agent_config = config.get_agent_config(metadata.agent_class)
-    agent_config.use_microagents = False
+    agent_config.enable_prompt_extensions = False
     return config
 
 

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/swe_bench/eval_infer.py

@@ -1,4 +1,6 @@
+import json
 import os
+import subprocess
 import tempfile
 import time
 from functools import partial
@@ -11,18 +13,23 @@ from swebench.harness.run_evaluation import (
 )
 from swebench.harness.test_spec import SWEbenchInstance, TestSpec, make_test_spec
 from swebench.harness.utils import load_swebench_dataset
+from tqdm import tqdm
 
+from evaluation.benchmarks.swe_bench.resource.mapping import (
+    get_instance_resource_factor,
+)
 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
@@ -66,7 +73,7 @@ def process_git_patch(patch):
     return patch
 
 
-def get_config(instance: pd.Series) -> AppConfig:
+def get_config(metadata: EvalMetadata, instance: pd.Series) -> AppConfig:
     # We use a different instance image for the each instance of swe-bench eval
     base_container_image = get_instance_docker_image(instance['instance_id'])
     logger.info(
@@ -74,18 +81,16 @@ def get_config(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=1800,
-            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,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,
@@ -123,7 +128,7 @@ def process_instance(
     else:
         logger.info(f'Starting evaluation for instance {instance.instance_id}.')
 
-    config = get_config(instance)
+    config = get_config(metadata, instance)
     instance_id = instance.instance_id
     model_patch = instance['model_patch']
     test_spec: TestSpec = instance['test_spec']
@@ -151,52 +156,52 @@ def process_instance(
     if runtime_failure_count > 0:
         config.sandbox.remote_runtime_resource_factor = min(
             config.sandbox.remote_runtime_resource_factor * (2**runtime_failure_count),
-            4,  # hardcode maximum resource factor to 4
+            8,
         )
         logger.warning(
-            f'This is the second attempt for instance {instance.instance_id}, setting resource factor to {config.sandbox.remote_runtime_resource_factor}'
+            f'This is the {runtime_failure_count + 1}th attempt for instance {instance.instance_id}, setting resource factor to {config.sandbox.remote_runtime_resource_factor}'
         )
 
-    runtime = create_runtime(config)
-    call_async_from_sync(runtime.connect)
-    # Get patch and save it to /tmp/patch.diff
-    with tempfile.TemporaryDirectory() as temp_dir:
-        # Patch file
-        patch_file_path = os.path.join(temp_dir, 'patch.diff')
-        with open(patch_file_path, 'w') as f:
-            f.write(model_patch)
-        runtime.copy_to(patch_file_path, '/tmp')
-        # Eval script
-        eval_script_path = os.path.join(temp_dir, 'eval.sh')
-        with open(eval_script_path, 'w') as f:
-            f.write(test_spec.eval_script)
-        runtime.copy_to(eval_script_path, '/tmp')
-
-    # Set +x
-    action = CmdRunAction(command='chmod +x /tmp/eval.sh')
-    action.timeout = 600
-    logger.info(action, extra={'msg_type': 'ACTION'})
-    obs = runtime.run_action(action)
-    logger.info(obs, extra={'msg_type': 'OBSERVATION'})
-    assert obs.exit_code == 0
-
-    # Apply patch
-    exec_command = (
-        'cd /testbed && '
-        "(git apply -v /tmp/patch.diff && echo 'APPLY_PATCH_PASS' || "
-        "(echo 'Failed to apply patch with git apply, trying with patch command...' && "
-        "(patch --batch --fuzz=5 -p1 -i /tmp/patch.diff && echo 'APPLY_PATCH_PASS' || "
-        "echo 'APPLY_PATCH_FAIL')))"
-    )
-    action = CmdRunAction(command=exec_command)
-    action.timeout = 600
-    obs = runtime.run_action(action)
-    assert isinstance(obs, CmdOutputObservation)
-    apply_patch_output = obs.content
-    assert isinstance(apply_patch_output, str)
-    instance['test_result']['apply_patch_output'] = apply_patch_output
-
     try:
+        runtime = create_runtime(config)
+        call_async_from_sync(runtime.connect)
+        # Get patch and save it to /tmp/patch.diff
+        with tempfile.TemporaryDirectory() as temp_dir:
+            # Patch file
+            patch_file_path = os.path.join(temp_dir, 'patch.diff')
+            with open(patch_file_path, 'w') as f:
+                f.write(model_patch)
+            runtime.copy_to(patch_file_path, '/tmp')
+            # Eval script
+            eval_script_path = os.path.join(temp_dir, 'eval.sh')
+            with open(eval_script_path, 'w') as f:
+                f.write(test_spec.eval_script)
+            runtime.copy_to(eval_script_path, '/tmp')
+
+        # Set +x
+        action = CmdRunAction(command='chmod +x /tmp/eval.sh')
+        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 obs.exit_code == 0
+
+        # Apply patch
+        exec_command = (
+            'cd /testbed && '
+            "(git apply -v /tmp/patch.diff && echo 'APPLY_PATCH_PASS' || "
+            "(echo 'Failed to apply patch with git apply, trying with patch command...' && "
+            "(patch --batch --fuzz=5 -p1 -i /tmp/patch.diff && echo 'APPLY_PATCH_PASS' || "
+            "echo 'APPLY_PATCH_FAIL')))"
+        )
+        action = CmdRunAction(command=exec_command)
+        action.set_hard_timeout(600)
+        obs = runtime.run_action(action)
+        assert isinstance(obs, CmdOutputObservation)
+        apply_patch_output = obs.content
+        assert isinstance(apply_patch_output, str)
+        instance['test_result']['apply_patch_output'] = apply_patch_output
+
         if 'APPLY_PATCH_FAIL' in apply_patch_output:
             logger.info(f'[{instance_id}] {APPLY_PATCH_FAIL}:\n{apply_patch_output}')
             instance['test_result']['report']['failed_apply_patch'] = True
@@ -212,7 +217,7 @@ def process_instance(
             # Run eval script in background and save output to log file
             log_file = '/tmp/eval_output.log'
             action = CmdRunAction(command=f'/tmp/eval.sh > {log_file} 2>&1 & echo $!')
-            action.timeout = 60  # Short timeout just to get the process ID
+            action.set_hard_timeout(300)  # Short timeout just to get the process ID
             obs = runtime.run_action(action)
 
             if isinstance(obs, CmdOutputObservation) and obs.exit_code == 0:
@@ -235,7 +240,7 @@ def process_instance(
                     check_action = CmdRunAction(
                         command=f'ps -p {pid} > /dev/null; echo $?'
                     )
-                    check_action.timeout = 60
+                    check_action.set_hard_timeout(300)
                     check_obs = runtime.run_action(check_action)
                     if (
                         isinstance(check_obs, CmdOutputObservation)
@@ -252,7 +257,7 @@ def process_instance(
 
                 # Read the log file
                 cat_action = CmdRunAction(command=f'cat {log_file}')
-                cat_action.timeout = 300
+                cat_action.set_hard_timeout(300)
                 cat_obs = runtime.run_action(cat_action)
 
                 # Grade answer
@@ -352,7 +357,14 @@ if __name__ == '__main__':
 
     # Load predictions
     assert args.input_file.endswith('.jsonl'), 'Input file must be a jsonl file.'
-    predictions = pd.read_json(args.input_file, lines=True)
+    required_fields = ['instance_id', 'model_patch', 'test_result']
+    with open(args.input_file) as f:
+        predictions = pd.DataFrame.from_records(
+            [
+                {k: v for k, v in json.loads(line).items() if k in required_fields}
+                for line in tqdm(f, desc='Loading predictions')
+            ]
+        )
     assert (
         'instance_id' in predictions.columns
     ), 'Input file must contain instance_id column.'
@@ -396,6 +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

@@ -0,0 +1,39 @@
+"""Mapping instance_id to resource_factor.
+
+Different instances may have different resource requirements.
+e.g., some instances may require more memory/CPU to run inference.
+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__))
+DEFAULT_RUNTIME_RESOURCE_FACTOR = int(
+    os.environ.get('DEFAULT_RUNTIME_RESOURCE_FACTOR', 1)
+)
+
+# dataset to resource mapping
+_global_resource_mapping: dict[str, dict[str, float]] = {}
+
+
+def get_resource_mapping(dataset_name: str) -> dict[str, float]:
+    if dataset_name not in _global_resource_mapping:
+        file_path = os.path.join(CUR_DIR, f'{dataset_name}.json')
+        if not os.path.exists(file_path):
+            logger.warning(f'Resource mapping for {dataset_name} not found.')
+            return None
+
+        with open(file_path, 'r') as f:
+            _global_resource_mapping[dataset_name] = json.load(f)
+        logger.info(f'Loaded resource mapping for {dataset_name}')
+    return _global_resource_mapping[dataset_name]
+
+
+def get_instance_resource_factor(dataset_name: str, instance_id: str) -> int:
+    resource_mapping = get_resource_mapping(dataset_name)
+    if resource_mapping is None:
+        return DEFAULT_RUNTIME_RESOURCE_FACTOR
+    return int(resource_mapping.get(instance_id, DEFAULT_RUNTIME_RESOURCE_FACTOR))

evaluation/benchmarks/swe_bench/run_infer.py

@@ -9,12 +9,16 @@ import toml
 from datasets import load_dataset
 
 import openhands.agenthub
+from evaluation.benchmarks.swe_bench.resource.mapping import (
+    get_instance_resource_factor,
+)
 from evaluation.utils.shared import (
     EvalException,
     EvalMetadata,
     EvalOutput,
     assert_and_raise,
     codeact_user_response,
+    get_default_sandbox_config_for_eval,
     get_metrics,
     is_fatal_evaluation_error,
     make_metadata,
@@ -27,7 +31,6 @@ from openhands.controller.state.state import State
 from openhands.core.config import (
     AgentConfig,
     AppConfig,
-    SandboxConfig,
     get_llm_config_arg,
     get_parser,
 )
@@ -41,9 +44,10 @@ from openhands.utils.async_utils import call_async_from_sync
 from openhands.utils.shutdown_listener import sleep_if_should_continue
 
 USE_HINT_TEXT = os.environ.get('USE_HINT_TEXT', 'false').lower() == 'true'
-USE_INSTANCE_IMAGE = os.environ.get('USE_INSTANCE_IMAGE', 'false').lower() == 'true'
+USE_INSTANCE_IMAGE = os.environ.get('USE_INSTANCE_IMAGE', 'true').lower() == 'true'
 RUN_WITH_BROWSING = os.environ.get('RUN_WITH_BROWSING', 'false').lower() == 'true'
 
+
 AGENT_CLS_TO_FAKE_USER_RESPONSE_FN = {
     'CodeActAgent': codeact_user_response,
 }
@@ -63,11 +67,11 @@ def get_instruction(instance: pd.Series, metadata: EvalMetadata):
         '<uploaded_files>\n'
         f'/workspace/{workspace_dir_name}\n'
         '</uploaded_files>\n'
-        f"I've uploaded a python code repository in the directory {workspace_dir_name}. Consider the following PR description:\n\n"
-        f'<pr_description>\n'
+        f"I've uploaded a python code repository in the directory {workspace_dir_name}. Consider the following issue description:\n\n"
+        f'<issue_description>\n'
         f'{instance.problem_statement}\n'
-        '</pr_description>\n\n'
-        'Can you help me implement the necessary changes to the repository so that the requirements specified in the <pr_description> are met?\n'
+        '</issue_description>\n\n'
+        'Can you help me implement the necessary changes to the repository so that the requirements specified in the <issue_description> are met?\n'
         "I've already taken care of all changes to any of the test files described in the <pr_description>. This means you DON'T have to modify the testing logic or any of the tests in any way!\n"
         'Your task is to make the minimal changes to non-tests files in the /workspace directory to ensure the <pr_description> is satisfied.\n'
         'Follow these steps to resolve the issue:\n'
@@ -118,24 +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,
-        ),
+        sandbox=sandbox_config,
         # do not mount workspace
         workspace_base=None,
         workspace_mount_path=None,
@@ -150,6 +153,7 @@ def get_config(
         codeact_enable_browsing=RUN_WITH_BROWSING,
         codeact_enable_llm_editor=False,
         condenser=metadata.condenser_config,
+        enable_prompt_extensions=False,
     )
     config.set_agent_config(agent_config)
     return config
@@ -173,7 +177,7 @@ def initialize_runtime(
     action = CmdRunAction(
         command=f"""echo 'export SWE_INSTANCE_ID={instance['instance_id']}' >> ~/.bashrc && echo 'export PIP_CACHE_DIR=~/.cache/pip' >> ~/.bashrc && echo "alias git='git --no-pager'" >> ~/.bashrc"""
     )
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -182,7 +186,7 @@ def initialize_runtime(
     )
 
     action = CmdRunAction(command="""export USER=$(whoami); echo USER=${USER} """)
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -194,7 +198,7 @@ def initialize_runtime(
 
         # inject the instance info
         action = CmdRunAction(command='mkdir -p /swe_util/eval_data/instances')
-        action.timeout = 600
+        action.set_hard_timeout(600)
         logger.info(action, extra={'msg_type': 'ACTION'})
         obs = runtime.run_action(action)
         logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -223,14 +227,14 @@ def initialize_runtime(
             '/swe_util/',
         )
         action = CmdRunAction(command='cat ~/.bashrc')
-        action.timeout = 600
+        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(obs.exit_code == 0, f'Failed to cat ~/.bashrc: {str(obs)}')
 
         action = CmdRunAction(command='source ~/.bashrc')
-        action.timeout = 600
+        action.set_hard_timeout(600)
         logger.info(action, extra={'msg_type': 'ACTION'})
         obs = runtime.run_action(action)
         logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -239,7 +243,7 @@ def initialize_runtime(
         assert_and_raise(obs.exit_code == 0, f'Failed to source ~/.bashrc: {str(obs)}')
 
         action = CmdRunAction(command='source /swe_util/instance_swe_entry.sh')
-        action.timeout = 3600
+        action.set_hard_timeout(600)
         logger.info(action, extra={'msg_type': 'ACTION'})
         obs = runtime.run_action(action)
         logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -249,7 +253,7 @@ def initialize_runtime(
         )
     else:
         action = CmdRunAction(command='source /swe_util/swe_entry.sh')
-        action.timeout = 1800
+        action.set_hard_timeout(1800)
         logger.info(action, extra={'msg_type': 'ACTION'})
         obs = runtime.run_action(action)
         logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -259,7 +263,7 @@ def initialize_runtime(
         )
 
     action = CmdRunAction(command=f'cd /workspace/{workspace_dir_name}')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -269,7 +273,7 @@ def initialize_runtime(
     )
 
     action = CmdRunAction(command='git reset --hard')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -278,14 +282,14 @@ def initialize_runtime(
     action = CmdRunAction(
         command='for remote_name in $(git remote); do git remote remove "${remote_name}"; done'
     )
-    action.timeout = 600
+    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(obs.exit_code == 0, f'Failed to remove git remotes: {str(obs)}')
 
     action = CmdRunAction(command='which python')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -316,17 +320,33 @@ def complete_runtime(
     workspace_dir_name = _get_swebench_workspace_dir_name(instance)
 
     action = CmdRunAction(command=f'cd /workspace/{workspace_dir_name}')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     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)}',
     )
 
     action = CmdRunAction(command='git config --global core.pager ""')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -336,7 +356,7 @@ def complete_runtime(
     )
 
     action = CmdRunAction(command='git add -A')
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -351,7 +371,7 @@ def complete_runtime(
         action = CmdRunAction(
             command=f'git diff --no-color --cached {instance["base_commit"]}'
         )
-        action.timeout = 600 + 100 * n_retries
+        action.set_hard_timeout(max(300 + 100 * n_retries, 600))
         logger.info(action, extra={'msg_type': 'ACTION'})
         obs = runtime.run_action(action)
         logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -399,7 +419,7 @@ def process_instance(
             8,
         )
         logger.warning(
-            f'This is the second attempt for instance {instance.instance_id}, setting resource factor to {config.sandbox.remote_runtime_resource_factor}'
+            f'This is the {runtime_failure_count + 1}th attempt for instance {instance.instance_id}, setting resource factor to {config.sandbox.remote_runtime_resource_factor}'
         )
     runtime = create_runtime(config)
     call_async_from_sync(runtime.connect)
@@ -479,6 +499,10 @@ def filter_dataset(dataset: pd.DataFrame, filter_column: str) -> pd.DataFrame:
                 subset = dataset[dataset[filter_column].isin(selected_ids)]
                 logger.info(f'Retained {subset.shape[0]} tasks after filtering')
                 return subset
+    skip_ids = os.environ.get('SKIP_IDS', '').split(',')
+    if len(skip_ids) > 0:
+        logger.info(f'Filtering {len(skip_ids)} tasks from "SKIP_IDS"...')
+        return dataset[~dataset[filter_column].isin(skip_ids)]
     return dataset
 
 
@@ -501,8 +525,10 @@ if __name__ == '__main__':
     # NOTE: It is preferable to load datasets from huggingface datasets and perform post-processing
     # so we don't need to manage file uploading to OpenHands's repo
     dataset = load_dataset(args.dataset, split=args.split)
-    logger.info(f'Loaded dataset {args.dataset} with split {args.split}')
     swe_bench_tests = filter_dataset(dataset.to_pandas(), 'instance_id')
+    logger.info(
+        f'Loaded dataset {args.dataset} with split {args.split}: {len(swe_bench_tests)} tasks'
+    )
 
     llm_config = None
     if args.llm_config:
@@ -531,6 +557,7 @@ if __name__ == '__main__':
     )
 
     output_file = os.path.join(metadata.eval_output_dir, 'output.jsonl')
+    print(f'### OUTPUT FILE: {output_file} ###')
     instances = prepare_dataset(swe_bench_tests, output_file, args.eval_n_limit)
 
     if len(instances) > 0 and not isinstance(

evaluation/benchmarks/swe_bench/scripts/eval/combine_final_completions.py

@@ -0,0 +1,69 @@
+import argparse
+import gzip
+import json
+import os
+from glob import glob
+
+from tqdm import tqdm
+
+tqdm.pandas()
+
+
+# Load trajectories for resolved instances
+def load_completions(output_dir: str, instance_id: str):
+    glob_path = os.path.join(output_dir, 'llm_completions', instance_id, '*.json')
+    files = sorted(glob(glob_path))  # this is ascending order
+    # pick the last file (last turn)
+    try:
+        file_path = files[-1]
+    except IndexError:
+        # print(f'No files found for instance {instance_id}: files={files}')
+        return None
+    with open(file_path, 'r') as f:
+        result = json.load(f)
+    # create messages
+    messages = result['messages']
+    messages.append(result['response']['choices'][0]['message'])
+    tools = result['kwargs']['tools']
+    return {
+        'messages': messages,
+        'tools': tools,
+    }
+
+
+parser = argparse.ArgumentParser()
+parser.add_argument('jsonl_path', type=str)
+args = parser.parse_args()
+
+output_dir = os.path.dirname(args.jsonl_path)
+output_path = os.path.join(output_dir, 'output.with_completions.jsonl.gz')
+
+# Check if output would be different from input
+needs_update = False
+with open(args.jsonl_path, 'r') as f_in:
+    for line in tqdm(f_in, desc='Checking for changes'):
+        data = json.loads(line)
+        new_completions = load_completions(output_dir, data['instance_id'])
+        current_completions = data.get('raw_completions')
+        if current_completions != new_completions:
+            needs_update = True
+            break
+
+if not needs_update:
+    print('No updates required. Skipping file update.')
+    exit(0)
+
+if os.path.exists(output_path):
+    print(f'Output file already exists at {output_path}, overwriting? (y/n)')
+    if input() != 'y':
+        print('Exiting...')
+        exit(0)
+
+# Process line by line
+with open(args.jsonl_path, 'r') as f_in, gzip.open(output_path, 'wt') as f_out:
+    for line in tqdm(f_in):
+        data = json.loads(line)
+        data['raw_completions'] = load_completions(output_dir, data['instance_id'])
+        f_out.write(json.dumps(data) + '\n')
+
+print(f'Saved compressed output to {output_path}')

evaluation/benchmarks/swe_bench/scripts/eval/compare_outputs.py

@@ -15,11 +15,25 @@ parser.add_argument(
     action='store_true',
     help='Show visualization paths for failed instances',
 )
+parser.add_argument(
+    '--only-x-instances',
+    action='store_true',
+    help='Only show instances that are ran by X',
+)
 args = parser.parse_args()
 
 df1 = pd.read_json(args.input_file_1, orient='records', lines=True)
 df2 = pd.read_json(args.input_file_2, orient='records', lines=True)
 
+if args.only_x_instances:
+    instance_ids_1 = set(df1['instance_id'].tolist())
+    print(
+        f'Before removing instances not in X={args.input_file_1}: Y={df2.shape[0]} instances'
+    )
+    df2 = df2[df2['instance_id'].isin(instance_ids_1)]
+    print(
+        f'After removing instances not in X={args.input_file_1}: Y={df2.shape[0]} instances'
+    )
 
 # Get the intersection of the instance_ids
 df = pd.merge(df1, df2, on='instance_id', how='inner')
@@ -86,7 +100,7 @@ repo_diffs = []
 for repo in all_repos:
     x_count = len(x_only_by_repo.get(repo, []))
     y_count = len(y_only_by_repo.get(repo, []))
-    diff = abs(x_count - y_count)
+    diff = y_count - x_count
     repo_diffs.append((repo, diff))
 
 # Sort by diff (descending) and then by repo name
@@ -106,7 +120,13 @@ for repo, diff in repo_diffs:
     repo_color = 'red' if is_significant else 'yellow'
 
     print(f"\n{colored(repo, repo_color, attrs=['bold'])}:")
-    print(colored(f'Difference: {diff} instances!', repo_color, attrs=['bold']))
+    print(
+        colored(
+            f'Difference: {diff} instances! (Larger diff = Y better)',
+            repo_color,
+            attrs=['bold'],
+        )
+    )
     print(colored(f'X resolved but Y failed: ({len(x_instances)} instances)', 'green'))
     if x_instances:
         print('  ' + str(x_instances))

evaluation/benchmarks/swe_bench/scripts/eval/convert_oh_output_to_md.py

@@ -4,6 +4,7 @@
 import argparse
 import json
 import os
+from glob import glob
 
 import pandas as pd
 from tqdm import tqdm
@@ -20,6 +21,7 @@ output_md_folder = args.oh_output_file.replace('.jsonl', '.viz')
 print(f'Converting {args.oh_output_file} to markdown files in {output_md_folder}')
 
 oh_format = pd.read_json(args.oh_output_file, orient='records', lines=True)
+output_dir = os.path.dirname(args.oh_output_file)
 
 swebench_eval_file = args.oh_output_file.replace('.jsonl', '.swebench_eval.jsonl')
 if os.path.exists(swebench_eval_file):
@@ -57,22 +59,172 @@ def convert_history_to_str(history):
     return ret
 
 
+# Load trajectories for resolved instances
+def load_completions(instance_id: str):
+    global output_dir
+    glob_path = os.path.join(output_dir, 'llm_completions', instance_id, '*.json')
+    files = sorted(glob(glob_path))  # this is ascending order
+    # pick the last file (last turn)
+    try:
+        file_path = files[-1]
+    except IndexError:
+        # print(f'No files found for instance {instance_id}: files={files}')
+        return None
+    with open(file_path, 'r') as f:
+        result = json.load(f)
+    # create messages
+    messages = result['messages']
+    messages.append(result['response']['choices'][0]['message'])
+    tools = result['kwargs']['tools']
+    return {
+        'messages': messages,
+        'tools': tools,
+    }
+
+
+def _convert_content(content) -> str:
+    ret = ''
+    if isinstance(content, list):
+        for item in content:
+            assert item['type'] == 'text', 'Only text is supported for now'
+            ret += f'{item["text"]}\n'
+    else:
+        assert isinstance(content, str), 'Only str is supported for now'
+        ret = content
+    return ret
+
+
+def convert_tool_call_to_string(tool_call: dict) -> str:
+    """Convert tool call to content in string format."""
+    if 'function' not in tool_call:
+        raise ValueError("Tool call must contain 'function' key.")
+    if 'id' not in tool_call:
+        raise ValueError("Tool call must contain 'id' key.")
+    if 'type' not in tool_call:
+        raise ValueError("Tool call must contain 'type' key.")
+    if tool_call['type'] != 'function':
+        raise ValueError("Tool call type must be 'function'.")
+
+    ret = f"<function={tool_call['function']['name']}>\n"
+    try:
+        args = json.loads(tool_call['function']['arguments'])
+    except json.JSONDecodeError as e:
+        raise ValueError(
+            f"Failed to parse arguments as JSON. Arguments: {tool_call['function']['arguments']}"
+        ) from e
+    for param_name, param_value in args.items():
+        is_multiline = isinstance(param_value, str) and '\n' in param_value
+        ret += f'<parameter={param_name}>'
+        if is_multiline:
+            ret += '\n'
+        ret += f'{param_value}'
+        if is_multiline:
+            ret += '\n'
+        ret += '</parameter>\n'
+    ret += '</function>'
+    return ret
+
+
+def format_traj(traj, first_n_turns=None, last_n_turns=None) -> str:
+    output = ''
+    system_message = None
+
+    # Handle system message if present
+    if traj[0]['role'] == 'system':
+        system_message = traj[0]
+        traj = traj[1:]
+        content = _convert_content(system_message['content'])
+        output += "*** System Message that describes the assistant's behavior ***\n"
+        output += f'{content}\n'
+
+    # Merge consecutive user messages first
+    merged_traj = []
+    current_messages = []
+
+    n_turns = len(traj)
+    for i, message in enumerate(traj):
+        # Skip this message if...
+        if (
+            # Case 1: first_n_turns specified and we're past it
+            (first_n_turns is not None and i >= first_n_turns and last_n_turns is None)
+            or
+            # Case 2: last_n_turns specified and we're before it
+            (
+                last_n_turns is not None
+                and i < n_turns - last_n_turns
+                and first_n_turns is None
+            )
+            or
+            # Case 3: both specified and we're in the middle section
+            (
+                first_n_turns is not None
+                and last_n_turns is not None
+                and i >= first_n_turns
+                and i < n_turns - last_n_turns
+            )
+        ):
+            continue
+
+        if message['role'] == 'user':
+            current_messages.append(message)
+        else:
+            if current_messages:
+                # Merge all accumulated user messages into one
+                merged_content = '\n'.join(
+                    _convert_content(msg['content']) for msg in current_messages
+                )
+                merged_traj.append({'role': 'user', 'content': merged_content})
+                current_messages = []
+            merged_traj.append(message)
+
+    # Don't forget to handle any remaining user messages
+    if current_messages:
+        merged_content = '\n'.join(
+            _convert_content(msg['content']) for msg in current_messages
+        )
+        merged_traj.append({'role': 'user', 'content': merged_content})
+
+    # Now process the merged trajectory
+    for i, message in enumerate(merged_traj):
+        role, content = message['role'], message['content']
+        content = _convert_content(content) if isinstance(content, list) else content
+        turn_id = i // 2 + 1
+        output += '-' * 100 + '\n'
+        output += f'*** Turn {turn_id} - {role.upper() if role != "tool" else "TOOL EXECUTION RESULT"} ***\n'
+
+        if role == 'user':
+            output += f'{content}\n'
+        elif role == 'tool':
+            output += f'{content}\n'
+        elif role == 'assistant':
+            output += f'{content}\n'
+            if (
+                'tool_calls' in message
+                and message['tool_calls'] is not None
+                and len(message['tool_calls']) > 0
+            ):
+                for toolcall_id, tool_call in enumerate(message['tool_calls']):
+                    output += f'### Tool Call {toolcall_id}\n'
+                    output += f'{convert_tool_call_to_string(tool_call)}\n'
+        else:
+            raise ValueError(f'Unexpected role: {role}')
+
+    output += '-' * 100 + '\n'
+    return output
+
+
 def write_row_to_md_file(row, instance_id_to_test_result):
     if 'git_patch' in row:
         model_patch = row['git_patch']
     elif 'test_result' in row and 'git_patch' in row['test_result']:
         model_patch = row['test_result']['git_patch']
     else:
-        raise ValueError(f'Row {row} does not have a git_patch')
+        print(f'Row {row} does not have a git_patch')
+        return
 
     test_output = None
-    if row['instance_id'] in instance_id_to_test_result:
-        report = instance_id_to_test_result[row['instance_id']].get('report', {})
-        resolved = report.get('resolved', False)
-        test_output = instance_id_to_test_result[row['instance_id']].get(
-            'test_output', None
-        )
-    elif 'report' in row and row['report'] is not None:
+    # Use result from output.jsonl FIRST if available.
+    if 'report' in row and row['report'] is not None:
         if not isinstance(row['report'], dict):
             resolved = None
             print(
@@ -80,6 +232,12 @@ def write_row_to_md_file(row, instance_id_to_test_result):
             )
         else:
             resolved = row['report'].get('resolved', False)
+    elif row['instance_id'] in instance_id_to_test_result:
+        report = instance_id_to_test_result[row['instance_id']].get('report', {})
+        resolved = report.get('resolved', False)
+        test_output = instance_id_to_test_result[row['instance_id']].get(
+            'test_output', None
+        )
     else:
         resolved = None
 
@@ -88,6 +246,8 @@ def write_row_to_md_file(row, instance_id_to_test_result):
     os.makedirs(output_md_folder, exist_ok=True)
     filepath = os.path.join(output_md_folder, filename)
 
+    completions = load_completions(instance_id)
+
     with open(filepath, 'w') as f:
         f.write(f'# {instance_id} (resolved: {resolved})\n')
 
@@ -97,7 +257,12 @@ def write_row_to_md_file(row, instance_id_to_test_result):
         f.write(json.dumps(row['metadata'], indent=2))
         f.write('\n```\n')
 
-        # Trajectory
+        # Completion
+        if completions is not None:
+            f.write('## Completion\n')
+            traj = completions['messages']
+            f.write(format_traj(traj))
+
         f.write('## History\n')
         f.write(convert_history_to_str(row['history']))
 

evaluation/benchmarks/swe_bench/scripts/eval/convert_oh_output_to_swe_json.py

@@ -22,7 +22,8 @@ def convert_row_to_swebench_format(row):
     elif 'test_result' in row and 'git_patch' in row['test_result']:
         model_patch = row['test_result']['git_patch']
     else:
-        raise ValueError(f'Row {row} does not have a git_patch')
+        print(f'WARNING: Row {row} does not have a git_patch')
+        model_patch = ''
 
     return {
         'instance_id': row['instance_id'],

evaluation/benchmarks/swe_bench/scripts/eval/update_output_with_eval.py

@@ -3,7 +3,7 @@ import json
 import os
 from collections import defaultdict
 
-import pandas as pd
+from tqdm import tqdm
 
 parser = argparse.ArgumentParser()
 parser.add_argument('input_file', type=str)
@@ -11,8 +11,7 @@ args = parser.parse_args()
 
 dirname = os.path.dirname(args.input_file)
 
-df = pd.read_json(args.input_file, lines=True)
-
+# Initialize counters and data structures
 instance_id_to_status = defaultdict(
     lambda: {
         'empty_generation': False,
@@ -23,15 +22,7 @@ instance_id_to_status = defaultdict(
     }
 )
 
-
-# Apply the status to the dataframe
-def apply_report(row):
-    instance_id = row['instance_id']
-    if instance_id in instance_id_to_status:
-        return dict(instance_id_to_status[instance_id])
-    return row.get('report', {})
-
-
+# Process official report if it exists
 swebench_official_report_json = os.path.join(dirname, 'report.json')
 openhands_remote_report_jsonl = args.input_file.replace(
     '.jsonl', '.swebench_eval.jsonl'
@@ -90,113 +81,159 @@ if os.path.exists(swebench_official_report_json):
             f'- [{instance_id}](./eval_outputs/{instance_id}/run_instance.log)\n'
         )
 
-    df['report'] = df.apply(apply_report, axis=1)
-
     with open(output_md_filepath, 'w') as f:
         f.write(output_md)
 
 elif os.path.exists(openhands_remote_report_jsonl):
     output_md_filepath = args.input_file.replace('.jsonl', '.swebench_eval.md')
 
-    df_eval = pd.read_json(openhands_remote_report_jsonl, lines=True, orient='records')
+    # First pass: Read eval report and count instances
+    instance_ids = set()
+    eval_instance_ids = set()
 
-    assert len(df['instance_id'].unique()) == len(
-        df
-    ), 'There are duplicate instance ids in the original output which is not allowed'
-    assert len(df_eval['instance_id'].unique()) == len(
-        df_eval
-    ), 'There are duplicate instance ids in the eval report which is not allowed'
+    # Count instances in original file
+    n_instances = 0
+    with open(args.input_file, 'r') as f:
+        for line in tqdm(f, desc='Counting instances in original file'):
+            data = json.loads(line)
+            instance_ids.add(data['instance_id'])
+            n_instances += 1
+    print(f'Total instances in original file: {n_instances}')
 
-    for _, row in df_eval.iterrows():
-        instance_id_to_status[row['instance_id']] = row['test_result']['report']
-    df['report'] = df.apply(apply_report, axis=1)
+    # Process eval report
+    n_eval_instances = 0
+    with open(openhands_remote_report_jsonl, 'r') as f:
+        for line in tqdm(f, desc='Processing eval report'):
+            data = json.loads(line)
+            instance_id = data['instance_id']
+            eval_instance_ids.add(instance_id)
+            n_eval_instances += 1
+            instance_id_to_status[instance_id] = data['test_result']['report']
+    print(f'Total instances in eval report: {n_eval_instances}')
 
-    report_is_dict = df['report'].apply(lambda x: isinstance(x, dict))
-    if not report_is_dict.all():
-        print(df[~report_is_dict])
-        raise ValueError(f'Report is not a dict, but a {type(row["report"])}')
+    # Verify no duplicates
+    assert (
+        len(instance_ids) == n_instances
+    ), 'Duplicate instance ids found in original output'
+    assert (
+        len(eval_instance_ids) == n_eval_instances
+    ), 'Duplicate instance ids found in eval report'
 
-    _n_instances = len(df)
-    _n_resolved = len(df[df['report'].apply(lambda x: x.get('resolved', False))])
-    _n_unresolved = _n_instances - _n_resolved
-    _n_empty_patch = len(
-        df[df['report'].apply(lambda x: x.get('empty_generation', False))]
-    )
-    _n_error = len(df[df['report'].apply(lambda x: x.get('error_eval', False))])
+    # Initialize counters
+    stats = {'total': len(instance_ids), 'resolved': 0, 'empty_patch': 0, 'error': 0}
+
+    # Collect instance IDs by category
+    resolved_ids = []
+    unresolved_ids = []
+    error_ids = []
+    empty_patch_ids = []
+    timeout_ids = []
+
+    # Process original file and categorize instances
+    with open(args.input_file, 'r') as f:
+        for line in f:
+            data = json.loads(line)
+            instance_id = data['instance_id']
+            report = instance_id_to_status[instance_id]
+
+            if report.get('resolved', False):
+                stats['resolved'] += 1
+                resolved_ids.append(instance_id)
+            else:
+                unresolved_ids.append(instance_id)
+
+            if report.get('empty_generation', False):
+                stats['empty_patch'] += 1
+                empty_patch_ids.append(instance_id)
+            if report.get('error_eval', False):
+                stats['error'] += 1
+                error_ids.append(instance_id)
+            if report.get('test_timeout', False):
+                timeout_ids.append(instance_id)
+
+    # Generate markdown report
+    def _instance_id_to_log_path(instance_id):
+        path = f"{args.input_file.replace('.jsonl', '.swebench_eval.logs')}/instance_{instance_id}.log"
+        return os.path.relpath(path, start=dirname)
+
+    # ... rest of markdown generation code remains the same ...
     output_md = (
         '# SWE-bench Report\n'
         'This folder contains the evaluation results of the SWE-bench using the [official evaluation docker containerization](https://github.com/princeton-nlp/SWE-bench/blob/main/docs/20240627_docker/README.md#choosing-the-right-cache_level).\n\n'
         '## Summary\n'
-        f'- submitted instances: {_n_instances}\n'
-        f'- empty patch instances: {_n_empty_patch}\n'
-        f'- resolved instances: {_n_resolved}\n'
-        f'- unresolved instances: {_n_unresolved}\n'
-        f'- error instances: {_n_error}\n'
+        f'- submitted instances: {stats["total"]}\n'
+        f'- empty patch instances: {stats["empty_patch"]}\n'
+        f'- resolved instances: {stats["resolved"]}\n'
+        f'- unresolved instances: {len(unresolved_ids)}\n'
+        f'- error instances: {stats["error"]}\n'
     )
 
-    def _instance_id_to_log_path(instance_id):
-        path = f"{args.input_file.replace('.jsonl', '.swebench_eval.logs')}/instance_{instance_id}.log"
-        # make it relative path
-        path = os.path.relpath(path, start=dirname)
-        return path
-
     output_md += '\n## Resolved Instances\n'
     # instance_id to status
-    for instance_id in sorted(
-        df[df['report'].apply(lambda x: x.get('resolved', False))][
-            'instance_id'
-        ].unique()
-    ):
+    for instance_id in resolved_ids:
         instance_id_to_status[instance_id]['resolved'] = True
         output_md += f'- [{instance_id}]({_instance_id_to_log_path(instance_id)})\n'
 
     output_md += '\n## Unresolved Instances\n'
-    for instance_id in sorted(
-        df[~df['report'].apply(lambda x: x.get('resolved', False))][
-            'instance_id'
-        ].unique()
-    ):
+    for instance_id in unresolved_ids:
         output_md += f'- [{instance_id}]({_instance_id_to_log_path(instance_id)})\n'
 
     output_md += '\n## Error Instances\n'
-    for instance_id in sorted(
-        df[df['report'].apply(lambda x: x.get('error_eval', False))][
-            'instance_id'
-        ].unique()
-    ):
+    for instance_id in error_ids:
         instance_id_to_status[instance_id]['error_eval'] = True
         output_md += f'- [{instance_id}]({_instance_id_to_log_path(instance_id)})\n'
 
     output_md += '\n## Empty Patch Instances\n'
-    for instance_id in sorted(
-        df[df['report'].apply(lambda x: x.get('empty_generation', False))][
-            'instance_id'
-        ].unique()
-    ):
+    for instance_id in empty_patch_ids:
         instance_id_to_status[instance_id]['empty_generation'] = True
         output_md += f'- [{instance_id}]({_instance_id_to_log_path(instance_id)})\n'
 
     output_md += '\n## Incomplete Instances\n'
-    for instance_id in sorted(
-        df[df['report'].apply(lambda x: x.get('test_timeout', False))][
-            'instance_id'
-        ].unique()
-    ):
+    for instance_id in timeout_ids:
         output_md += f'- [{instance_id}]({_instance_id_to_log_path(instance_id)})\n'
+
     with open(output_md_filepath, 'w') as f:
         f.write(output_md)
+
 else:
     print(
         f'No report file found: Both {swebench_official_report_json} and {openhands_remote_report_jsonl} do not exist.'
     )
     exit()
 
+# Before backup and update, check if any changes would be made
+needs_update = False
+with open(args.input_file, 'r') as infile:
+    for line in tqdm(infile, desc='Checking for changes'):
+        data = json.loads(line)
+        instance_id = data['instance_id']
+        current_report = data.get('report', {})
+        new_report = instance_id_to_status[
+            instance_id
+        ]  # if no report, it's not resolved
+        if current_report != new_report:
+            needs_update = True
+            break
+
+if not needs_update:
+    print('No updates detected. Skipping file update.')
+    exit()
+
+# Backup and update the original file row by row
 if os.path.exists(args.input_file + '.bak'):
     conf = input('Existing backup file found. Do you want to overwrite it? (y/n)')
     if conf != 'y':
         exit()
     os.remove(args.input_file + '.bak')
 
-# backup the original file
 os.rename(args.input_file, args.input_file + '.bak')
-df.to_json(args.input_file, orient='records', lines=True)
+
+# Process and write file row by row
+with open(args.input_file + '.bak', 'r') as infile, open(
+    args.input_file, 'w'
+) as outfile:
+    for line in tqdm(infile, desc='Updating output file'):
+        data = json.loads(line)
+        instance_id = data['instance_id']
+        data['report'] = instance_id_to_status[instance_id]
+        outfile.write(json.dumps(data) + '\n')

evaluation/benchmarks/swe_bench/scripts/eval_infer.sh

@@ -76,7 +76,7 @@ echo "Running SWE-bench evaluation"
 echo "=============================================================="
 
 RUN_ID=$(date +"%Y%m%d_%H%M%S")
-N_PROCESS=16
+N_PROCESS=4
 
 if [ -z "$INSTANCE_ID" ]; then
     echo "Running SWE-bench evaluation on the whole input file..."
@@ -87,7 +87,7 @@ if [ -z "$INSTANCE_ID" ]; then
         --dataset_name "$DATASET_NAME" \
         --split "$SPLIT" \
         --predictions_path $SWEBENCH_FORMAT_JSONL \
-        --timeout 1800 \
+        --timeout 3600 \
         --cache_level instance \
         --max_workers $N_PROCESS \
         --run_id $RUN_ID
@@ -133,7 +133,7 @@ else
         --dataset_name "$DATASET_NAME" \
         --split "$SPLIT" \
         --predictions_path $SWEBENCH_FORMAT_JSONL \
-        --timeout 1800 \
+        --timeout 3600 \
         --instance_ids $INSTANCE_ID \
         --cache_level instance \
         --max_workers $N_PROCESS \

evaluation/benchmarks/swe_bench/scripts/run_infer.sh

@@ -108,7 +108,14 @@ if [ -z "$N_RUNS" ]; then
   echo "N_RUNS not specified, use default $N_RUNS"
 fi
 
+# Skip runs if the run number is in the SKIP_RUNS list
+# read from env variable SKIP_RUNS as a comma separated list of run numbers
+SKIP_RUNS=(${SKIP_RUNS//,/ })
 for i in $(seq 1 $N_RUNS); do
+  if [[ " ${SKIP_RUNS[@]} " =~ " $i " ]]; then
+    echo "Skipping run $i"
+    continue
+  fi
   current_eval_note="$EVAL_NOTE-run_$i"
   echo "EVAL_NOTE: $current_eval_note"
   run_eval $current_eval_note

evaluation/benchmarks/the_agent_company/README.md

@@ -17,11 +17,14 @@ When the `run_infer.sh` script is started, it will automatically pull all task i
 
 ```bash
 ./evaluation/benchmarks/the_agent_company/scripts/run_infer.sh \
-  --agent-llm-config <agent-llm-config>  \
-  --env-llm-config <env-llm-config> \
-  --outputs-path <outputs-path> \
-  --server-hostname <server-hostname> \
-  --version <version>
+  --agent-llm-config <agent-llm-config, default to 'agent'>  \
+  --env-llm-config <env-llm-config, default to 'env'> \
+  --outputs-path <outputs-path, default to outputs> \
+  --server-hostname <server-hostname, default to localhost> \
+  --version <version, default to 1.0.0> \
+  --start-percentile <integer from 0 to 99, default to 0> \
+  --end-percentile <integer from 1 to 100, default to 100>
+
 
 # Example
 ./evaluation/benchmarks/the_agent_company/scripts/run_infer.sh \
@@ -29,7 +32,9 @@ When the `run_infer.sh` script is started, it will automatically pull all task i
   --env-llm-config claude-3-5-sonnet-20240620 \
   --outputs-path outputs \
   --server-hostname localhost \
-  --version 1.0.0
+  --version 1.0.0 \
+  --start-percentile 10 \
+  --end-percentile 20
 ```
 
 - `agent-llm-config`: the config name for the agent LLM. This should match the config name in config.toml. This is the LLM used by the agent (e.g. CodeActAgent).
@@ -37,7 +42,11 @@ When the `run_infer.sh` script is started, it will automatically pull all task i
 - `outputs-path`: the path to save trajectories and evaluation results.
 - `server-hostname`: the hostname of the server that hosts all the web services. It could be localhost if you are running the evaluation and services on the same machine. If the services are hosted on a remote machine, you must use the hostname of the remote machine rather than IP address.
 - `version`: the version of the task images to use. Currently, the only supported version is 1.0.0.
+- `start-percentile`: the start percentile of the task split, must be an integer between 0 to 99.
+- `end-percentile`: the end percentile of the task split, must be an integer between 1 to 100 and larger than start-percentile.
 
-The script is idempotent. If you run it again, it will resume from the last checkpoint. It would usually take a few days to finish evaluation.
+The script is idempotent. If you run it again, it will resume from the last checkpoint. It would usually take 2 days to finish evaluation if you run the whole task set.
+To speed up evaluation, you can use `start-percentile` and `end-percentile` to split the tasks for higher parallelism,
+provided concurrent runs are **targeting different servers**.
 
 Note: the script will automatically skip a task if it encounters an error. This usually happens when the OpenHands runtime dies due to some unexpected errors. This means even if the script finishes, it might not have evaluated all tasks. You can manually resume the evaluation by running the script again.

evaluation/benchmarks/the_agent_company/browsing.py

@@ -262,12 +262,14 @@ def pre_login(
             instruction = action.to_instruction()
 
             browser_action = BrowseInteractiveAction(browser_actions=instruction)
-            browser_action.timeout = 10000
+            browser_action.set_hard_timeout(10000)
             logger.info(browser_action, extra={'msg_type': 'ACTION'})
             obs: BrowserOutputObservation = runtime.run_action(browser_action)
             logger.debug(obs, extra={'msg_type': 'OBSERVATION'})
             if save_screenshots:
-                image_data = base64.b64decode(obs.screenshot)
+                image_data = base64.b64decode(
+                    obs.screenshot.replace('data:image/png;base64,', '')
+                )
                 with open(os.path.join(directory, f'{image_id}.png'), 'wb') as file:
                     file.write(image_data)
                     image_id += 1

evaluation/benchmarks/the_agent_company/run_infer.py

@@ -13,14 +13,16 @@ from typing import List
 import yaml
 from browsing import pre_login
 
+from evaluation.utils.shared import get_default_sandbox_config_for_eval
 from openhands.controller.state.state import State
 from openhands.core.config import (
     AppConfig,
     LLMConfig,
-    SandboxConfig,
+    get_agent_config_arg,
     get_llm_config_arg,
     get_parser,
 )
+from openhands.core.config.agent_config import AgentConfig
 from openhands.core.logger import openhands_logger as logger
 from openhands.core.main import create_runtime, run_controller
 from openhands.events.action import CmdRunAction, MessageAction
@@ -34,7 +36,10 @@ def get_config(
     task_short_name: str,
     mount_path_on_host: str,
     llm_config: LLMConfig,
+    agent_config: AgentConfig | None,
 ) -> AppConfig:
+    sandbox_config = get_default_sandbox_config_for_eval()
+    sandbox_config.base_container_image = base_container_image
     config = AppConfig(
         run_as_openhands=False,
         max_budget_per_task=4,
@@ -42,21 +47,21 @@ def get_config(
         save_trajectory_path=os.path.join(
             mount_path_on_host, f'traj_{task_short_name}.json'
         ),
-        sandbox=SandboxConfig(
-            base_container_image=base_container_image,
-            enable_auto_lint=True,
-            # using host network to access the host machine from the container
-            use_host_network=True,
-            # large enough timeout, since some testcases take very long to run
-            timeout=300,
-            api_key=os.environ.get('ALLHANDS_API_KEY', None),
-        ),
+        sandbox=sandbox_config,
         # we mount trajectories path so that trajectories, generated by OpenHands
         # controller, can be accessible to the evaluator file in the runtime container
         workspace_mount_path=mount_path_on_host,
         workspace_mount_path_in_sandbox='/outputs',
     )
     config.set_llm_config(llm_config)
+    if agent_config:
+        config.set_agent_config(agent_config)
+    else:
+        logger.info('Agent config not provided, using default settings')
+        agent_config = AgentConfig(
+            enable_prompt_extensions=False,
+        )
+        config.set_agent_config(agent_config)
     return config
 
 
@@ -80,13 +85,13 @@ def load_dependencies(runtime: Runtime) -> List[str]:
 def init_task_env(runtime: Runtime, hostname: str, env_llm_config: LLMConfig):
     command = (
         f'SERVER_HOSTNAME={hostname} '
-        f'LITELLM_API_KEY={env_llm_config.api_key} '
+        f'LITELLM_API_KEY={env_llm_config.api_key.get_secret_value() if env_llm_config.api_key else None} '
         f'LITELLM_BASE_URL={env_llm_config.base_url} '
         f'LITELLM_MODEL={env_llm_config.model} '
         'bash /utils/init.sh'
     )
     action = CmdRunAction(command=command)
-    action.timeout = 900
+    action.set_hard_timeout(900)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -147,11 +152,21 @@ def run_solver(
         os.makedirs(screenshots_dir, exist_ok=True)
         for image_id, obs in enumerate(state.history):
             if isinstance(obs, BrowserOutputObservation):
-                image_data = base64.b64decode(obs.screenshot)
+                image_data = base64.b64decode(
+                    obs.screenshot.replace('data:image/png;base64,', '')
+                )
                 with open(
                     os.path.join(screenshots_dir, f'{image_id}.png'), 'wb'
                 ) as file:
                     file.write(image_data)
+                if obs.set_of_marks:
+                    som_image_data = base64.b64decode(
+                        obs.set_of_marks.replace('data:image/png;base64,', '')
+                    )
+                    with open(
+                        os.path.join(screenshots_dir, f'{image_id}_som.png'), 'wb'
+                    ) as file:
+                        file.write(som_image_data)
 
     if save_final_state:
         os.makedirs(state_dir, exist_ok=True)
@@ -165,14 +180,14 @@ def run_evaluator(
     runtime: Runtime, env_llm_config: LLMConfig, trajectory_path: str, result_path: str
 ):
     command = (
-        f'LITELLM_API_KEY={env_llm_config.api_key} '
+        f'LITELLM_API_KEY={env_llm_config.api_key.get_secret_value() if env_llm_config.api_key else None} '
         f'LITELLM_BASE_URL={env_llm_config.base_url} '
         f'LITELLM_MODEL={env_llm_config.model} '
         f"DECRYPTION_KEY='theagentcompany is all you need' "  # Hardcoded Key
         f'python_default /utils/eval.py --trajectory_path {trajectory_path} --result_path {result_path}'
     )
     action = CmdRunAction(command=command)
-    action.timeout = 600
+    action.set_hard_timeout(600)
     logger.info(action, extra={'msg_type': 'ACTION'})
     obs = runtime.run_action(action)
     logger.info(obs, extra={'msg_type': 'OBSERVATION'})
@@ -214,6 +229,10 @@ if __name__ == '__main__':
     )
     args, _ = parser.parse_known_args()
 
+    agent_config: AgentConfig | None = None
+    if args.agent_config:
+        agent_config = get_agent_config_arg(args.agent_config)
+
     agent_llm_config: LLMConfig | None = None
     if args.agent_llm_config:
         agent_llm_config = get_llm_config_arg(args.agent_llm_config)
@@ -254,7 +273,7 @@ if __name__ == '__main__':
     else:
         temp_dir = tempfile.mkdtemp()
     config: AppConfig = get_config(
-        args.task_image_name, task_short_name, temp_dir, agent_llm_config
+        args.task_image_name, task_short_name, temp_dir, agent_llm_config, agent_config
     )
     runtime: Runtime = create_runtime(config)
     call_async_from_sync(runtime.connect)