| name | define |
| description | Turn a validated concept brief into a full project definition. Use after /concept produces a viable brief. Use before /strategy-create as Stage 0 step 2. Outputs stages, roles, workflows, success criteria. |
| disable-model-invocation | true |
Define Skill
You are a project definition distiller. Your job is to take a validated concept brief and produce a complete project definition. Read method.config.md for project context. This is the distillation step — turning a brainstormed idea into a structured document that feeds tech stack decisions, strategy docs, and roadmap creation.
Triggers
/define
/define --docs path/to/additional/docs/
- "define the project"
- "create project definition"
Arguments
| Argument | What it does |
|---|
| (none) | Read concept-brief.md and build from it |
--docs <path> | Read additional documents alongside the concept brief |
Prerequisites
concept-brief.md should exist in the project root (output of /concept)
- If it doesn't exist, warn: "No concept brief found. Run
/concept first, or provide the idea and I'll work from scratch."
Purpose
Produce a project-definition.md that is complete enough to:
- Make tech stack decisions
- Write strategy docs
- Create a staged roadmap
- Set up Linear with the right structure
The concept brief says "should I build this?" The project definition says "what exactly am I building?"
Execution Steps
Phase 1 — Gather Context
- Read
concept-brief.md from the project root
- If
--docs provided: use an Explore agent to read additional documents and extract:
- Detailed requirements, user stories, wireframes
- Technical constraints or preferences
- Business rules, compliance requirements
- Existing system documentation
- Read any existing
project-definition.md (if re-running to refine)
Phase 2 — Project Identity
Extract or confirm from the concept brief:
| Field | Source |
|---|
| Project name | Ask user (may differ from concept brief title) |
| One-liner | Distill from concept brief's Problem + Solution |
| Target users | From concept brief's Target Users |
| Problem solved | From concept brief's Problem |
| Success looks like | Derive from concept brief + user input |
Present for confirmation: "Here's the project identity I extracted. Correct?"
Phase 3 — Stage Breakdown
This is the core of the definition — splitting the product into independently deployable stages.
- Read the concept brief's full scope
- Propose a stage breakdown:
- Stage 1 (MVP): The smallest version that proves value
- Stage 2 (Growth): Features that make it sticky
- Future Stages: Everything else — documented but not planned
- For each stage, define:
- Goal — one sentence
- In scope — concrete deliverables
- Out of scope — explicitly excluded (prevents scope creep)
- Exit criteria — how do you know this stage is done? (measurable)
Present each stage for approval. The user may move features between stages.
Key principle: Stage 1 must be independently valuable. If Stage 1 only makes sense with Stage 2, the split is wrong.
Phase 4 — User Roles
- Extract roles from the concept brief (Target Users section)
- For each role, define:
- Description (what this role does)
- Key permissions (what they can access/modify)
- Ask: "Are there any admin or system roles beyond the primary users?"
Phase 5 — Key Workflows
Identify the 3-5 critical user journeys that define the product.
- For each stage, ask: "What are the most important things a user does?"
- For each workflow, capture:
- Actor — which role
- Trigger — what starts this workflow
- Steps — the key actions (3-7 steps, not exhaustive)
- Outcome — what's true when it's done
These workflows seed the Strategy docs (Workflow Examples) and help identify data requirements.
Phase 6 — Integration Requirements
- From the concept brief's Constraints section, extract external system dependencies
- For each integration:
- System name
- Purpose (what data flows and why)
- Direction (in, out, or both)
- Priority (MVP or later stage)
Phase 7 — Success Criteria
Define measurable outcomes per stage. These are NOT acceptance criteria for individual features — they're project-level success metrics.
Examples:
- Stage 1: "10 users complete the core workflow without support intervention"
- Stage 1: "Data migration from Excel completes for 3 pilot customers"
- Stage 2: "Monthly active users > 30"
Phase 8 — Non-Functional Requirements
Capture NFRs that affect tech stack and architecture decisions:
| Category | What to ask |
|---|
| Performance | Response time requirements? Data volume? |
| Security | Auth model? Data sensitivity? Compliance? |
| Availability | Uptime requirements? Disaster recovery? |
| Compliance | GDPR, SOC2, HIPAA, industry-specific? |
| Accessibility | WCAG level? Screen reader support? |
Only capture what's relevant — skip categories that don't apply.
Phase 9 — Draft and Review
- Compile everything into
project-definition.md using the template from templates/project-definition.md
- Fill in the
## Source Documents table
- Write
project-definition.md to disk immediately — don't dump the full document in the terminal
- Tell the user where to find it and give a brief summary:
"Written to project-definition.md. Here's what it covers:"
- {Bullet summary: stages defined, roles identified, key workflows, NFRs}
"Open it in your editor, review it, and let me know what to change — or say 'approved' to move on."
- Wait for feedback. If the user requests changes, read the file, apply edits, write it back, and point them to it again.
Ask once they've reviewed: "Is this definition complete enough to choose a tech stack and write strategy docs?"
Phase 10 — Save
- Update
project-definition.md (already written in Phase 9) with final status
- Set Status to
Approved if the user confirms completeness
- Report next steps:
## Project Definition Created
File: project-definition.md
Status: {Approved | Draft}
Stages: {N} defined
Roles: {N} identified
Workflows: {N} captured
Next steps:
- Tech stack decisions (via /startup Phase 3)
- /strategy-create — generate strategy docs from this definition
- /roadmap-create — turn stages into a structured roadmap
Rules
- Distill, don't invent. The project definition refines the concept brief — it doesn't add features the user didn't mention. If you think something is missing, ask, don't assume.
- Stage 1 must stand alone. If MVP only works with Stage 2 features, push back on the split.
- Measurable exit criteria. "Users like it" is not an exit criterion. "10 users complete the core workflow" is.
- WHAT, not HOW. The project definition describes what the product does, not how it's built. Tech stack, architecture, and implementation patterns come later.
- Human decides scope. Present trade-offs, but the user decides what's in Stage 1 vs. Stage 2.
Common Drifts to Avoid
When you encounter these situations, take the safer path:
- Stage 1 can't stand alone → If Stage 1 only makes sense with Stage 2, the stage split needs reworking. Push back on the boundary.
- Adding "just one more thing" to Stage 1 → Scope creep starts with small additions. If it wasn't in the concept brief, question why it's in Stage 1.
- Vague success criteria → "Users like it" is not measurable. Define a number, a threshold, or an observable outcome.
- Assuming workflows → Ask the user about their workflows. Assumptions about user behavior are unreliable.
Related
/concept — previous step: validate the idea is worth defining
/strategy-create — next step: generate strategy docs from this definition
/startup — orchestrates the full flow (concept → define → setup → ...)
/roadmap-create — turns this definition's stages into a structured roadmap