Flows

A Flow is the fundamental unit of work in DevFlow. Every feature, bug fix, or maintenance task is tracked as a Flow. Unlike simple tickets, Flows have a structured lifecycle with planning, approval, implementation, and review phases — ensuring quality at every step.

Flow Types

Each Flow has a type that determines how it appears on the Kanban board and how it is prioritized:

• Feature — Standard work items for new functionality, improvements, or changes. Shown with a blue indicator.
• Hotfix — Urgent fixes that need immediate attention. Shown with a red indicator.

You can change the type while the Flow is still in the idea state.

Creating a Flow

1. Open your project
2. Click “New Flow” or use the + button on the Kanban board
3. The Flow is created immediately in the idea state with the title “Untitled”
4. Edit the title by clicking on it
5. Add a description explaining what needs to be done
6. Add acceptance criteria — specific conditions that define when the work is complete
7. Choose the type (Feature or Hotfix)

Clear acceptance criteria are especially important when working with AI agents, as they guide the planning and implementation process.

Flow Detail View

Clicking on a Flow opens its detail page. The content shown depends on the Flow’s current state, but you always have access to:

• Title — Click to edit inline at any time
• Display ID — A short identifier like MA-3 that you can copy to clipboard
• Flow ID — The internal ID, also copyable
• Claude command — A pre-formatted command you can copy and paste into your AI agent to start working on this Flow
• Type indicator — Feature or Hotfix
• Assignee — Who is responsible for this Flow
• Release badge — Which release this Flow belongs to (if any)
• Work Here button — Start your timer on this Flow

State-Specific Content

As a Flow moves through its lifecycle, the detail page shows different views:

• Idea — Description editor and acceptance criteria list
• Planning — The AI agent’s implementation plan as it is being created
• Approval — The completed plan for you to review, with approve/reject options
• Ready — The approved plan, waiting for implementation to start
• In Progress — Implementation progress with task tracking and git activity
• Review — Agent summary, testing instructions, and links to the PR and commits
• Done — Final summary of everything that was accomplished

Additional Sections

• Tasks — Sub-items that break the Flow into smaller steps. Tasks are created during planning and tracked during implementation.
• Git Information — Branch name, commit history, and pull request links
• Agent Sessions — A timeline showing when AI agents worked on this Flow, what they did, and how long it took
• Time Entries — Time blocks linked to this Flow from the timer

Assigning Flows

You can assign a Flow to any team member in the project. Click the assignee selector in the Flow header to choose a person. Assignees can also be changed directly from the Kanban board card.

Unassigned Flows are visible to everyone and can be picked up by any team member.

Assigning Flows to Releases

Flows can be linked to a release to group related work together. When a Flow is part of a release, a release badge appears in the Flow header, and the Flow is visible when filtering the Kanban board by that release.

Release assignment can happen automatically if you have configured auto-assignment in the project rules.

Flow Lifecycle

Every Flow moves through seven states in three phases:

Phase	States	What happens
Preparation	idea, planning, approval	The work is defined, planned, and the plan is reviewed
Development	ready, in_progress, review	Code is written, tested, and reviewed
Complete	done	Work is finished and merged

The key principle: you decide, the AI executes. You approve plans before code is written. You review results before they are merged.

For a complete reference of all states, transitions, and what is required at each step, see the Flow States reference.