OneDev 15.1 enhances and better integrates existing AI capabilities, enabling a more streamlined development workflow.
Tod no longer operates as an MCP server. Instead, it works as a CLI that interacts directly with the OneDev server, accompanied by a collection of skills that streamline development workflows within coding agents.
For example, you can use prompts such as:
For more information, see:
https://code.onedev.io/onedev/tod/~files/main/readme.md
In addition to opening workspaces on branches, you can now open workspaces directly on issues and pull requests.
When a workspace is opened for an issue, the issue is automatically transitioned to the In Progress state when using the default issue workflow.
Tod CLI and skills are automatically installed and configured in workspace images, allowing you to simply instruct your coding agent with prompts such as:
Besides running workspaces on the OneDev server, Enterprise Edition can run workspaces as Kubernetes pods with files stored on persistent volume claims (PVCs), or on remote agents.
This helps offload resource-intensive development tasks from the server and improves server stability in team environments.
When you ask AI-assisted codebase questions from the dashboard or project pages, the AI can perform regex and symbol searches across all projects you have access to, enabling it to answer questions that span multiple repositories.
The AI reviewer can now post inline code comments for issues it identifies during pull request reviews.
As code changes are made or feedback is provided on those comments, the AI continuously re-evaluates its findings and automatically determines whether comments should be resolved or re-raised.
For a complete list of improvements and bug fixes, see: