Tasks

A SaaS-focused workflow to turn business requirements into delivery plans, tasks, and progress tracking.

Use this simple loop to ship your SaaS faster with AI: define what you need, get a plan, execute tasks, and track progress.

Minimal Markdown Tracker

This is a minimal task tracker based on markdown files. For more serious and advanced workflows, you can follow:

File Structure

PROJECT.md

plan.md

PROGRESS.md

task-1.md

PROJECT.md: Your main project brief and requirements. Update this to define your core customer and success criteria.
plan.md: The AI-generated delivery plan ordered by business impact and dependencies.
PROGRESS.md: Tracked execution progress and delivered outcomes, appended after completing tasks.
tasks/backlog/: Pending tasks sliced from plan.md. The highest priority ones are worked on first.
tasks/in-progress/: The single task currently being implemented by the AI. Only one task goes here at a time.
tasks/done/: Completed task files are moved here for reference.

Use the following direct commands in agent chat to manage and execute your SaaS roadmap automatically.

Turn your rough idea into an actionable roadmap.

Write your ideas in .agents/tasks/PROJECT.md.
Send the /tasks-plan command in your chat.
The AI reads PROJECT.md, evaluates business value, outputs plan.md, and slices the plan into actionable .md files directly in tasks/backlog/.

Execute items one by one. Run this each time you want the AI to ship the next feature or fix.

Send the /tasks-execute command.
The AI identifies the top-priority task in your backlog/ and moves it to in-progress/.
It implements the code end-to-end to deliver the feature or fix.
When done, the AI moves the task to done/ and records the impact in PROGRESS.md.

Repeat /tasks-execute until your backlog is clear.