V
Viban
Get Started

Boards & Tasks

Learn how to organize your work with Viban's Kanban system.

Understanding Boards

A board in Viban represents a project or workspace. Each board contains:

  • Columns: Visual lanes for different task states
  • Tasks: Individual work items that move through columns
  • Repository: Connected Git repository for code changes

Default Columns

Every board comes with four default columns:

Column Purpose
Todo Tasks waiting to be started
In Progress Tasks currently being worked on by AI
In Review Completed tasks awaiting review
Done Finished and merged tasks

Working with Tasks

Creating Tasks

  1. Click "Add Task" in the Todo column
  2. Enter a title (required)
  3. Add a description (optional but recommended)
  4. Save the task

Task Descriptions

Write clear, actionable descriptions for best AI results:

Good Example:

Add user authentication with email/password login.

Requirements:
- Registration form with email validation
- Login form with remember me option
- Password reset via email
- Session management with JWT tokens

Poor Example:

Add login

Refining Tasks

Use the "Refine" button to automatically enhance task descriptions:

  1. Create a task with a basic description
  2. Click the "Refine" button
  3. AI transforms it into a detailed prompt with:
    • Clear objectives
    • Acceptance criteria
    • Scope boundaries
    • Implementation suggestions

Task Position

Tasks don't have explicit status fields - their status is determined by which column they belong to. Moving a task between columns changes its effective state. Tasks also have an agent_status that tracks AI execution state:

Agent Status Description
idle No AI activity
thinking AI is processing
executing AI is running commands
error An error occurred

Moving Tasks

Drag and Drop

Simply drag a task card to move it between columns. When you drag to "In Progress":

  1. Viban creates a git worktree
  2. Launches the AI agent
  3. Starts streaming output

Automatic Transitions

  • In Progress → In Review: When AI completes work
  • In Review → Done: When you merge the changes

Task Output

While a task is in progress, you can view:

  • Real-time output: Live streaming from the AI agent
  • Git changes: Files modified in the worktree
  • Execution logs: Detailed execution history

Best Practices

  1. One feature per task: Keep tasks focused and atomic
  2. Use refinement: Let AI enhance your descriptions
  3. Review carefully: Always review AI-generated code
  4. Test locally: Run tests before merging