Browser client for monitoring
AI Models & Assistants / Knowledge library
Agent execution runtime
Executes tool calls and tracks a model-driven workflow.
Why learn Agent execution runtime?
Makes file edits and command execution auditable and controllable.
The problem it solves
Connects model proposals to permissioned actions in an environment.
Where it came from
Agent execution runtime: Executes tool calls and tracks a model-driven workflow.
Build your understanding
Work through these exercises in order. Try the task before opening a reference.
- Explain what Agent execution runtime does.
- Identify a use case for Agent execution runtime.
- Describe a limitation of Agent execution runtime.
- Name a compatible tool for Agent execution runtime.
- Compare Agent execution runtime against an alternative.
Documentation and sources
- Primary source / model card (docs.openhands.dev)
Catalog research notes
- See the AI Workbench for source dates, pricing evidence and configuration details.
Connect the ideas
Built on
Real World Example
Automated Legacy Data Migration Orchestration
Financial institutions managing decades of fragmented mainframe records.
The business problem
Manual data extraction and transformation processes are prone to human error and high latency, leading to inconsistent database synchronization across legacy systems.
How this topic solves it
Utilizing the OpenHands Agent Server to execute tool-driven workflows that automate the extraction, validation, and migration of data records within isolated, secure sandbox environments.
- Initialize the Agent Server to expose the remote execution API for secure task management.
- Define custom tools within the Software Agent SDK to interface with legacy database schemas.
- Configure the Sandbox Server to provide isolated environments for each migration batch.
- Deploy the Agent Canvas to monitor the model-driven workflow and verify tool call outputs.
- Execute the migration sequence via the Agent Server API to ensure consistent state transitions.
The flow shows how the Agent Server manages tools and sandboxes to process data.
Orchestrates execution and tools
Isolated execution environment
Source of migration data
- Agent Canvas → Sends commands → Agent Server
- Agent Server → Spawns execution → Sandbox Server
- Sandbox Server → Extracts records → Legacy Database
Expected business outcome
A reliable, repeatable migration pipeline that reduces manual intervention while maintaining strict data integrity through isolated execution.
Limitations and checks
The primary limitation is the overhead of managing isolated sandbox environments for complex tasks; this can be verified by monitoring the latency of the Sandbox Server during high-concurrency migration batches.