plan-project

name	plan-project
description	This skill should be used when the user asks to "plan a project", "brainstorm a project", "plan a new app", "decompose a project into stories", "greenfield project planning", "break down an application idea", or wants to go from an application description to a set of Khanrad issues and stories.
version	1.0.0

Khanrad Plan Project Skill

This skill takes a greenfield application description, guides the user through iterative refinement, decomposes the application into domains, generates stories and epics using parallel subagents, and populates a Khanrad board with tagged, prioritized issues.

When This Skill Applies

• User has an application idea and wants to plan it out
• User wants to break down a project into implementable stories
• User asks to brainstorm, decompose, or plan a greenfield project
• User wants to go from a description to a Khanrad board full of issues
• User mentions "greenfield", "new project", "project planning", or "story generation"

Key Concepts

• Domain — a bounded context or feature area within the application (e.g., auth, payments, notifications). Each domain becomes a tag prefix: domain:<name>.
• Tag taxonomy — issues are tagged with structured labels for filtering:
  • domain:<name> — feature area (e.g., domain:auth, domain:orders)
  • phase:mvp / phase:v2 / phase:future — release phase
  • type:epic / type:story / type:spike — work type
  • cross-cutting — items spanning multiple domains
• Parallel subagents — each domain's stories are generated by an independent subagent running concurrently, then collected for user review.
• Refinement loop — the skill asks iterative questions to build a complete picture before generating stories. It never generates stories from a vague description alone.

Workflow Overview

1. Intake — collect the application description and ask refining questions about users, workflows, constraints, integrations, and MVP scope.
2. Decompose — identify domains/bounded contexts and confirm with the user.
3. Generate — spawn parallel subagents (one per domain) to produce epics and stories.
4. Review — present stories grouped by domain, let the user refine, adjust priorities, and confirm the MVP cut.
5. Populate — create a Khanrad project and board, then create all approved issues with tags, priorities, and state organization.

Output

A fully populated Khanrad board with issues organized by domain tags, release phase, priority, and board state (Backlog for MVP, Ice Box for future work).