Init

Route Based on User Intent

After presenting current state, determine what the user wants:

• "add app" or user wants to register a new app → jump to Add-App Sub-Flow (bottom of this skill)
  • Idea graduation: If marketing/ideas/ exists and contains a validated idea (frontmatter go_no_go: go), ask: "You have a validated idea — [idea_name]. Want to use it as the starting point for this app?" If yes, pre-populate Step 2 fields from the idea file: problem hypothesis → one-liner, ICP hypothesis → positioning basics, competitors → top competitors, pricing hypothesis → pricing model. Update the idea file status to graduated.
• "update info" or user wants to change app details → read the relevant file, ask what to update, write changes. Do NOT recreate the directory structure.
• "reset" followed by a section name → confirm which section to reset, then re-populate only that section from templates. Never reset without listing exactly what will be overwritten and getting confirmation.
• No specific request → recommend the most appropriate next skill based on current phase and status (use the logic from Step 5)

Re-Initialization Safety

If the user asks to "reinitialize" or "start over":

1. NEVER overwrite without explicit confirmation
2. List all files that would be affected and what data would be lost
3. Offer alternatives: "Would you rather update specific information, or add a new app?"
4. Only proceed with full reset if the user explicitly confirms after seeing the impact

Step 1: Portfolio Information

Gather brand-level information. Keep this conversational and quick.

Questions to Ask

Ask these questions in a natural conversational flow (not a rigid form):

1. Brand name: "What's your brand or company name?" (e.g., "Mentilead")
2. App count: "How many Shopify apps do you have — just one, or multiple?"
3. Monthly budget: "Roughly what's your monthly marketing budget?"
   • Present tiers: $0 (bootstrapping), $1-200, $200-500, $500-1000, $1000+

Single-App Language Rule

If the user has only ONE app:

• Suppress all portfolio terminology throughout the rest of this flow
• Never use "portfolio", "brand-level", or "cross-sell" language
• Treat the app name as the brand name unless the user specifies otherwise
• The experience should feel like setting up marketing for ONE app, not a portfolio

If the user has MULTIPLE apps:

• Use portfolio language naturally
• Explain that shared strategy (positioning, content, partnerships) lives at the brand level
• Per-app files (listing, keywords, funnel, reviews) are maintained separately

Step 2: App Information

Gather information about each app. This is the most detailed step — be conversational, not robotic.

Core Information (All Apps)

For each app, gather:

1. App name: The name as it appears (or will appear) on the Shopify App Store
2. App slug: Auto-generate from name (e.g., "B2B Onboard" → b2b-onboard). Confirm with user.
3. One-liner: "In one sentence, what does this app do for merchants?"
4. Category: Ask the user to pick or describe. Common categories:
   • B2B / Wholesale
   • Marketing / Email / SMS
   • Analytics / Reporting
   • Operations / Fulfillment
   • Customer Service / Support
   • Payments / Checkout
   • Store Design / UX
   • Inventory / Product Management
   • SEO / Content
   • Subscriptions / Loyalty
5. Current status: Where is the app right now?
   • pre-launch — Still building, not on the App Store yet
   • beta — On the App Store, actively seeking early users
   • launched — Live and accepting installs
   • growing — Established with active user base, optimizing for growth
   • scaling — Significant traction, expanding channels and partnerships

Conditional Information (Launched / Growing / Scaling Only)

Only ask these if the app's status is launched, growing, or scaling:

6. Pricing model: Free, freemium, paid-only, or usage-based? What are the price points?
7. Current installs: Approximate active install count
8. Reviews: How many reviews? What's the average rating?
9. MRR: Current monthly recurring revenue (approximate is fine)
10. Shopify plan requirement: Minimum plan merchants need (Basic / Shopify / Advanced / Plus)

For pre-launch and beta apps, skip questions 6-10 entirely. Set these fields to defaults:

• Pricing: "Not finalized"
• Installs: 0
• Reviews: 0
• Rating: N/A
• MRR: $0
• Shopify plan: "Basic" (default)

Positioning Basics (All Apps)

11. Key differentiators: "What makes your app different from competitors? List your top 2-3 unique strengths."
12. Top competitors: "Who are your top 1-3 competitors on the Shopify App Store?" (name or URL)

Multi-App Loop

If the user has multiple apps:

• After gathering info for the first app, ask: "Ready to add the next app, or would you prefer to add more later?"
• If "add later": proceed with the apps gathered so far. They can use the add-app flow anytime.
• If "add next": repeat Step 2 for the next app.

Step 3: Directory Creation & Template Population

Create the full marketing/ directory structure and populate files with gathered data.

Directory Tree to Create

marketing/
├── MEMORY.md
├── STATUS.md
├── logs/
│   └── .gitkeep
├── positioning/
│   ├── brand.md
│   ├── icp.md
│   └── jtbd.md
├── partnerships/
│   ├── pipeline.md
│   └── templates.md
├── content/
│   ├── strategy.md
│   └── ideas.md
├── outreach/
│   ├── templates.md
│   ├── targets.md
│   └── campaigns/
│       └── .gitkeep
├── experiments/
│   ├── backlog.md
│   └── .gitkeep
├── experiment/           (if experiment layer enabled)
│   ├── observations.md
│   ├── autonomy-log.md
│   ├── signals.md
│   └── drafts/
│       └── .gitkeep
├── metrics/
│   ├── portfolio.md
│   └── weekly/
│       └── .gitkeep
└── apps/
    └── {slug}/           (one per app)
        ├── app.md
        ├── listing.md
        ├── keywords.md
        ├── positioning.md
        ├── funnel.md
        ├── reviews.md
        └── experiments/
            └── .gitkeep

Template Sources

Read each template from the plugin's templates/ directory, populate with gathered data, and write to the target path:

Experiment Layer (Optional)

After creating the standard directory structure, ask:

"Are you using GrowthOS to document a public experiment? (yes/no)"

If yes:

1. Create marketing/experiment/ directory with drafts/ subdirectory (.gitkeep)
2. Copy templates/observations.md → marketing/experiment/observations.md
3. Copy templates/autonomy-log.md → marketing/experiment/autonomy-log.md
4. Copy templates/signals.md → marketing/experiment/signals.md
5. Set experiment_chapter: "Chapter 1: The Wager" in marketing/STATUS.md frontmatter
6. Read references/experiment-thesis.md and add to marketing/MEMORY.md:

   ## Experiment Layer
   - **Thesis:** {thesis from experiment-thesis.md}
   - **Current Chapter:** Chapter 1: The Wager
   - **Observations logged:** 0
   - **Autonomy Score:** Not yet measured
   

If no: skip. No experiment files created.

COO Configuration (Optional)

After creating the directory structure (and optionally the experiment layer), ask: "Do you want to configure the COO agent? (recommended)"

If yes, ask:

1. "How many hours per week do you have for this work?" -> write to references/coo-config.md Time Budget
2. "What are your working days?" -> write to Preferred working days
3. "What is the current development phase for each app?" -> write to Development Priorities and set dev_status: {app}: {phase} in marketing/STATUS.md frontmatter
4. "What are your publishing commitments?" -> write with defaults (Substack monthly, LinkedIn weekly)

Write answers to references/coo-config.md. Add last_coo_briefing: never to marketing/STATUS.md frontmatter. Create marketing/experiment/coo-log.md with header: "# COO Briefing Log\n\nDecisions and follow-ups from daily COO briefings."

If no: skip. User can configure later.

Files Created Empty (Just Headers or .gitkeep)

These files get placeholder content — they'll be populated by dedicated skills later:

• marketing/positioning/icp.md — Header only: "# Ideal Customer Profile\n\nRun the Positioning skill to develop your ICP."
• marketing/positioning/jtbd.md — Header only: "# Jobs-to-be-Done Research\n\nRun the Positioning skill to start JTBD interviews."
• marketing/partnerships/pipeline.md — Header only: "# Partnership Pipeline\n\nRun the Partnerships skill to build your pipeline."
• marketing/partnerships/templates.md — Header only: "# Partnership Templates\n\nRun the Partnerships skill to generate templates."
• marketing/content/strategy.md — Header only: "# Content Strategy\n\nRun the Content Strategy skill to plan your content."
• marketing/content/ideas.md — Header only: "# Content Ideas\n\nRun the Content Strategy skill to generate ideas."
• marketing/outreach/templates.md — Header only: "# Outreach Templates\n\nRun the Outreach skill to create email templates."
• marketing/outreach/targets.md — Header only: "# Outreach Targets\n\nRun the Outreach skill to define target segments."
• marketing/apps/{slug}/keywords.md — Header only: "# Keyword Research — {App Name}\n\nRun the ASO Optimizer skill to research keywords."
• marketing/apps/{slug}/positioning.md — Header only: "# Positioning — {App Name}\n\nRun the Positioning skill to develop app-specific positioning."
• marketing/apps/{slug}/reviews.md — Header only: "# Reviews — {App Name}\n\nRun the Review Manager skill to build your review strategy."

Population Rules

When populating templates with gathered data:

• Replace all YYYY-MM-DD date placeholders with today's date
• Replace [Brand Name] and [App Name] with actual names
• Fill frontmatter fields with gathered data
• For fields without gathered data, leave the template default/placeholder
• For single-app users: replace portfolio-level language in MEMORY.md heading with "Marketing Memory — {App Name}" instead of "Marketing Memory — {Brand Name} Portfolio"

Step 4: Quick Positioning Assessment

Provide immediate value by generating a brief positioning assessment. This is NOT a full positioning exercise (that's the Positioning skill) — just enough to give the developer actionable direction.

Reference Files to Read

Read these two files to inform the assessment:

• references/merchant-segments.md — for segmentation frameworks and ICP patterns
• references/funnel-benchmarks.md — for realistic milestone expectations and conversion rates

Generate Assessment

Based on the gathered app information (category, status, competitors, differentiators), generate:

1. Competitive Landscape Summary (3-5 sentences)

• Where does this app sit relative to the named competitors?
• What market category does it compete in?
• Is the category crowded or underserved?
• What's the app's likely competitive position? (challenger, niche player, innovator)

2. Initial ICP Hypothesis (3-5 bullet points)

Using the segmentation data from merchant-segments.md, suggest:

• Most likely Shopify plan segment for this app's merchants
• Most likely merchant vertical(s)
• Likely technical sophistication level
• Likely business stage
• Key pain point this app probably addresses

Frame this as a hypothesis to validate, not a final answer:

"Based on your category and competitors, your initial ICP is likely: [hypothesis]. The Positioning skill will help you validate and refine this."

3. Recommended First Actions (Exactly 3)

Based on the app's current status, recommend 3 specific first actions:

Pre-launch apps:

1. Complete ICP and positioning with the Positioning skill
2. Draft your App Store listing copy with the ASO Optimizer
3. Identify 5 potential beta testers in Shopify Community forums

Beta apps:

1. Reach out to 10 merchants who use a competing app (Outreach skill)
2. Complete your App Store listing optimization (ASO Optimizer)
3. Ask your first 3 active users for honest feedback (Review Manager)

Launched apps (few reviews):

1. Optimize your App Store listing — title and first 3 screenshots are critical (ASO Optimizer)
2. Set up a review acquisition strategy (Review Manager) — you need 20-60 reviews for organic discovery
3. Create your first experiment to improve install rate (Experiment Engine)

Growing apps:

1. Start an experiment to improve your weakest funnel stage (Experiment Engine)
2. Build a content strategy to create an organic acquisition channel (Content Strategy)
3. Identify 3 complementary apps for cross-promotion (Partnerships)

Write Assessment

Write the competitive landscape summary and ICP hypothesis to:

• marketing/positioning/positioning.md — append to the Competitive Context section

Do NOT overwrite the positioning template structure. Add the assessment content within the existing template sections.

Step 5: Next Skill Recommendation

Based on the app's status, recommend the most impactful next skill to use.

Recommendation Logic

Pre-launch:

"Your most impactful next step is defining your positioning and ICP. This determines everything — your listing copy, outreach targets, and content strategy." Recommended: Positioning skill — "Tell me to work on positioning" or "Help me define my ICP"

Beta:

"You need early users and feedback fast. Focus on direct outreach to merchants who already use competing apps, and start collecting honest reviews." Recommended: Outreach skill + Review Manager — "Help me find beta users" or "Help me get reviews"

Launched (few reviews, < 20):

"Your App Store listing is your #1 conversion tool — 70%+ of installs come from in-store search. Optimize it, then focus on getting to 20+ reviews for organic discovery." Recommended: ASO Optimizer + Review Manager — "Help me optimize my listing" or "Help me get more reviews"

Growing (20+ reviews):

"You have a validated product. Now it's time to experiment systematically and build scalable acquisition channels." Recommended: Experiment Engine + Content Strategy — "Help me run experiments" or "Help me plan content"

Scaling:

"Focus on partnerships for leverage and portfolio-wide optimization." Recommended: Partnerships + Metrics Dashboard — "Help me find partners" or "Show me my metrics"

Multi-App Recommendation

For multi-app portfolios, base the recommendation on the app with the EARLIEST status. A portfolio with one pre-launch and one growing app should focus on the pre-launch app first.

Session Log & State Update

After completing initialization, update persistent state.

Create Session Log

Create marketing/logs/{today}.md where {today} is today's date in YYYY-MM-DD format:

# Session Log — {today}

## GrowthOS Initialized

- **Brand:** {brand_name}
- **Apps registered:** {count}
{for each app:}
  - {app_name} ({slug}) — {status}
{end}
- **Monthly budget:** {budget_tier}
- **Initial assessment:** Generated competitive landscape and ICP hypothesis
- **Recommended next step:** {recommended skill from Step 5}
{if experiment layer enabled:}
- **Experiment layer:** Enabled — Chapter 1: The Wager
{if COO configured:}
- **COO agent:** Configured — {hours}hrs/week, {working_days}

Update STATUS.md

Set initial priorities in STATUS.md based on the recommended next skill:

• current_phase: Set to the earliest app status across all apps
• active_app: For single-app, set to the app slug. For multi-app, set to the earliest-status app slug.
• This Week's Priorities: Populate with the 3 recommended first actions from Step 4
• Upcoming Actions: Add "Run {recommended skill}" as the first action

Confirmation Message

After everything is created, present a summary to the user:

GrowthOS is set up! Here's what was created:

marketing/
├── MEMORY.md — Your marketing knowledge base
├── STATUS.md — Your marketing dashboard
├── positioning/ — Brand and app positioning
├── apps/{slug}/ — {app_name} files (listing, funnel, reviews)
├── experiments/ — Experiment backlog
├── metrics/ — Portfolio metrics tracker
├── content/ — Content strategy (empty, ready for Content skill)
├── outreach/ — Outreach templates (empty, ready for Outreach skill)
├── partnerships/ — Partnership pipeline (empty, ready for Partnerships skill)
{if experiment layer enabled:}
├── experiment/       — Observation log, autonomy tracker, signals, drafts
{if COO configured:}
├── references/coo-config.md — COO operational preferences
└── logs/{today}.md — Today's session log

{Quick positioning assessment from Step 4}

Next step: {recommendation from Step 5}

Add-App Sub-Flow

Used when marketing/ already exists and the user wants to add a new app.

1. Gather app info: Run through Step 2 questions for the new app only (core info, conditional metrics, positioning basics)
2. Create app directory: Create marketing/apps/{new-slug}/ with all 7 files (app.md, listing.md, keywords.md, positioning.md, funnel.md, reviews.md, experiments/.gitkeep)
3. Update MEMORY.md:
   • Add the new app to the apps list in frontmatter
   • Add a row to the "Apps in Portfolio" table
4. Update STATUS.md:
   • Add a row to the "Per-App Overview" table
   • Set active_app to the new app slug
5. Positioning assessment: Run Step 4 for the new app only
6. Session log: Append to today's log (or create new if none exists):

## App Added: {app_name}

- **Slug:** {slug}
- **Status:** {status}
- **Assessment:** {brief competitive landscape note}

Language Update

If this is the second app being added (transitioning from single to multi-app):

• Update MEMORY.md heading from "Marketing Memory — {App Name}" to "Marketing Memory — {Brand Name} Portfolio"
• Inform the user: "Since you now have multiple apps, your marketing memory and dashboard now show portfolio-level views. Each app's specific files are in marketing/apps/{slug}/."

Field Mapping Table