Is your enhancement related to a problem? Please describe
Developers can run AI coding tasks in Che, but there is no single place to manage multiple agents working on related tasks. It is difficult to see what is running, return to previous tasks, and review the resulting changes.
Describe the solution you'd like
Provide an Agentic Workspace MVP where a developer can:
- Create or select an Agentic Workspace to organize one or more agent tasks.
- Start one or more agent sessions, assign tasks, and see their status.
- Use the terminal interface provided by each agent. The Agentic Workspace should not recreate a separate chat UI for Codex, Claude Code, OpenCode, or other supported agents.
- Use the Changes right panel to review files and diffs produced by the selected agent.
- Use the Terminal right panel to access the Agentic Workspace development environment, inspect files, and run commands.
- Use Open in VS Code to open the Agentic Workspace in the Dev Spaces development environment through VS Code.
The Agentic Workspace should build on the agent runtime support described in Eclipse Che issue #23921 and che-operator PR #2184 as its backend and execution foundation.
An Agentic Workspace groups the agent sessions and development environment used for a piece of software. It uses the Dev Spaces development environment when the developer needs to inspect, test, or continue the work.
The MVP should support this flow:
- Create or select an Agentic Workspace.
- Start a task through the agent's (Claude/Codex/OpenCode) existing TUI.
- Leave and return to see whether the task is running, blocked, or complete.
- Review the changes and use the Terminal panel when needed.
- Open the Agentic Workspace in VS Code to continue development.
This MVP does not include a new agent chat UI or replace the terminal interfaces provided by individual agents. Implementation and technology choices are intentionally open.
View the running Agentic Workspace MVP
Describe alternatives you've considered
No response
Additional context
No response
Is your enhancement related to a problem? Please describe
Developers can run AI coding tasks in Che, but there is no single place to manage multiple agents working on related tasks. It is difficult to see what is running, return to previous tasks, and review the resulting changes.
Describe the solution you'd like
Provide an Agentic Workspace MVP where a developer can:
The Agentic Workspace should build on the agent runtime support described in Eclipse Che issue #23921 and che-operator PR #2184 as its backend and execution foundation.
An Agentic Workspace groups the agent sessions and development environment used for a piece of software. It uses the Dev Spaces development environment when the developer needs to inspect, test, or continue the work.
The MVP should support this flow:
This MVP does not include a new agent chat UI or replace the terminal interfaces provided by individual agents. Implementation and technology choices are intentionally open.
View the running Agentic Workspace MVP
Describe alternatives you've considered
No response
Additional context
No response