Columns
| Column | Status | Description |
|---|---|---|
| Todo | todo | Tasks waiting to be worked on |
| In Progress | in_progress | Tasks currently being executed by an AI agent |
| Done | done | Successfully completed tasks |
| Cancelled | cancelled | Cancelled or failed tasks |
Task Cards
Each task card displays:- Title — The task description
- Identifier — Team-scoped number (e.g.,
ENG-1) - Priority — Color-coded indicator (low, medium, high)
- Labels — Colored badges
- PR link — Clickable link when execution produces a PR
- Execution progress — Progress bar during execution
- Selection checkbox — For batch operations
Task Card Actions
Right-click or use the action menu on any card:| Action | Available when |
|---|---|
| Move to In Progress | Task is in Todo |
| Execute | Task is in In Progress |
| Cancel | Task is executing |
| Delete | Any status |
| View Details | Any status |
Task Creation
Click + Add task in any column to open the creation form:| Field | Required | Description |
|---|---|---|
| Title | Yes | Task description (what you want built) |
| Description | No | Additional context and requirements |
| Priority | No | Low, Medium (default), or High |
| Labels | No | Categorization tags |
| Team | No | Assign to a team for scoped numbering |
| Project | No | Associate with a project |
| Due date | No | Target completion date |
Task Detail View
Click any task card to open the detail panel showing:- Full task metadata (title, description, priority, labels, dates)
- Execution logs (real-time during execution)
- Execution progress and elapsed time
- PR URL and outcome summary
- Pending permission requests from the AI agent
- Edit and delete controls
Selection and Batch Operations
Enter selection mode
Click the selection toggle on a column header to enable checkboxes on task cards.
Model Selector
The model selector dropdown in the project config panel lets you choose which AI model to use for execution. It shows:- All connected providers with their available models
- Free models from OpenCode (Minimix, GLM5, KimiK2.5)
- Model capabilities (reasoning, tool calling)
- Cost per token (input/output)
Real-Time Updates
The board receives real-time updates via Server-Sent Events:- Task creation, updates, and deletion from other views
- Execution progress during agent runs
- Batch status changes
- Permission requests from the AI agent

