Canvas-to-Portfolio Bootstrap

Seed a cogni-portfolio project from a Lean Canvas (or Business Model Canvas). This skill bridges the gap between a founding-stage hypothesis document and the structured entity model that the rest of the portfolio pipeline expects.

Why This Exists

A lean canvas is a hypothesis document — it captures what you think the business is. The portfolio pipeline is a messaging system — it captures what you sell. For a founding-stage company these are the same thing, but the data shapes are different. This skill does the translation so that downstream skills (propositions, solutions, packages, export) work without modification.

The key insight: a canvas already contains products, markets, and pricing buried in its sections — they just need to be extracted, properly typed, and marked as early-stage (maturity: "concept", readiness: "planned", priority: "beachhead").

Prerequisites

• A lean canvas file (markdown, typically with numbered sections: Problem, Customer Segments, UVP, Solution, Channels, Revenue Streams, Cost Structure, Key Metrics, Unfair Advantage)
• No existing portfolio project required — this skill can create one via portfolio-setup if needed, or enrich an existing project

Canvas Section → Portfolio Entity Mapping

Canvas Section	Portfolio Entity	How It Maps
Customer Segments	Markets	Each segment becomes a market. Primary → priority: "beachhead", Secondary → "expansion", Tertiary → "aspirational". Pain points feed downstream proposition DOES/MEANS.
Solution (layers/components)	Products	Each distinct solution layer or component becomes a product with maturity: "concept" and inferred revenue_model.
Solution (capabilities within layers)	Features	Specific capabilities mentioned within solution layers become features with readiness: "planned".
Revenue Streams	Solution pricing templates	Pricing assumptions seed solution pricing or subscription fields. The revenue type (license, subscription, consulting, maintenance) determines solution_type.
Problem	Context	Pain points inform proposition DOES statements downstream — stored as canvas_context.problems in portfolio.json for reference.
UVP	Context	The overarching value proposition informs product positioning — stored as canvas_context.uvp in portfolio.json.
Channels	Context	Distribution channels are not portfolio entities but inform market descriptions — stored as canvas_context.channels in portfolio.json.
Cost Structure	Context	Cost categories inform solution cost_model downstream — stored as canvas_context.cost_structure in portfolio.json.
Key Metrics	Context	Stored as canvas_context.key_metrics in portfolio.json for reference.
Unfair Advantage	Context	Feeds differentiators in portfolio-context export — stored as canvas_context.unfair_advantage in portfolio.json.
Assumptions to Validate	Context	Stored as canvas_context.assumptions in portfolio.json. Referenced during proposition and solution creation so hypotheses stay visible.

Sections that don't map directly to entities are preserved as canvas_context in portfolio.json. This matters because downstream skills (especially propositions and solutions) can read this context to stay grounded in the founding-stage reality rather than generating messaging as if the company were established.

Workflow

1. Locate and Parse the Canvas

Read the canvas file provided by the user. Support common lean canvas formats:

• Numbered sections (## 1. Problem, ## 2. Customer Segments, etc.)
• Named sections (## Problem, ## Customer Segments, etc.)
• Bullet-point or prose within sections

Extract content from each section into a structured intermediate representation. If a section is empty or marked with "?" (common for early drafts), note it as unfilled.

2. Check for Existing Portfolio Project

Search for portfolio.json under a cogni-portfolio/ path in the workspace. Two paths:

• No project exists: Proceed to Step 3 to create one.
• Project exists: Ask the user whether to enrich the existing project with canvas data or start fresh. If enriching, cross-reference extracted entities against existing products/, features/, and markets/ directories to avoid duplicates.

3. Bootstrap Project (if needed)

If no portfolio project exists, extract setup fields from the canvas:

• Company name: May need to ask the user (canvases rarely include this)
• Description: Derive from the UVP section
• Industry: Infer from the solution domain, confirm with user
• Products: Extract from the Solution section

Run the setup scaffold (same as portfolio-setup):

bash $CLAUDE_PLUGIN_ROOT/scripts/project-init.sh "<workspace-dir>" "<project-slug>"

Write portfolio.json with company context plus the canvas_context object.

4. Present Extraction Plan

Before creating any entities, present the full extraction plan as a table:

Products (from Solution section)

Slug	Name	Revenue Model	Maturity	Source
iwae-platform	IWAE Platform	subscription	concept	Solution Layer 1
plugin-marketplace	Plugin Marketplace	hybrid	concept	Solution Layer 2
consulting-services	Implementation Consulting	project	concept	Solution Layer 3

Features (from Solution capabilities)

Slug	Product	Name	Readiness	Source
claude-obsidian-integration	iwae-platform	Claude + Obsidian Integration	planned	Solution 1, bullet 1
gdpr-deployment	iwae-platform	GDPR-Compliant EU Deployment	planned	Solution 1, bullet 3

Markets (from Customer Segments)

Slug	Name	Priority	Region	Source
boutique-consultants-dach	Capacity-Constrained Consulting Boutiques	beachhead	dach	Segment: Primary
ai-partners-dach	AI Portfolio-Expanding Consultancies	expansion	dach	Segment: Secondary
sme-sales-dach	AI-Ambitious SME Sales Functions	aspirational	dach	Segment: Tertiary

Canvas Context (preserved in portfolio.json)

Section	Status	Content Summary
Problem	filled	10h/week automation potential...
UVP	filled	World's first open-source IWAE...
Channels	filled	4 channels identified
Revenue Streams	filled	5 streams with pricing assumptions
Cost Structure	filled	5 cost categories
Key Metrics	unfilled	—
Unfair Advantage	unfilled	—
Assumptions	filled	5 assumptions to validate

Allow the user to:

• Approve all — create everything as proposed
• Edit — modify slugs, names, assignments, or skip individual entities
• Adjust region — the default region is dach; the user may want eu, global, or another region

5. Infer Revenue Models and Solution Types

Map canvas revenue streams to portfolio solution types. This inference drives how downstream solutions are structured:

Revenue Stream Pattern	revenue_model	solution_type
License, SaaS, subscription, per-seat, per-month	subscription	subscription
Consulting, implementation, project-based	project	project
Certification, training, onboarding	project	project
Maintenance, support, retainer	subscription	subscription
Partnership, revenue-share, referral	partnership	partnership
Mixed (e.g., SaaS + consulting)	hybrid	hybrid

When pricing assumptions exist in the canvas (e.g., "€15k/year", "€3k/participant"), preserve them as canvas_pricing annotations on the relevant product. These become seed values when the user later runs the solutions skill.

6. Write Entity Files

For each confirmed entity, write JSON following the schemas in $CLAUDE_PLUGIN_ROOT/references/data-model.md.

Products — write to products/{slug}.json:

• Set maturity: "concept" (founding stage)
• Set revenue_model based on Step 5 inference
• Set positioning from the UVP section if applicable to this product
• Include "source_file": "<canvas-filename>" for traceability

Features — write to features/{slug}.json:

• Set readiness: "planned" (not yet built) unless the canvas indicates otherwise
• Set product_slug to the parent product
• Draft a purpose field (5-12 words): a customer-readable statement answering "what is this feature FOR?" — the problem it solves or the capability it provides. Purpose sits between name (label) and description (mechanism). Example: name "Cloud Monitoring", purpose "Real-time visibility into cloud health and incidents".
• Write market-independent IS-layer descriptions (what it IS, not what it does for a specific segment)
• Assign sort_order following the value-to-utility spectrum: customer-facing value features get low numbers (10, 20, 30...), infrastructure/utility features get high numbers (70+). Use increments of 10.

Markets — write to markets/{slug}.json:

• Set priority based on segment tier (primary/secondary/tertiary)
• Include pain points from the canvas in description
• Leave tam/sam/som empty with a note — founding-stage businesses typically have hypotheses, not validated sizing. The markets skill can add sizing later.
• Set segmentation fields where the canvas provides enough detail (employee range, vertical, etc.)

7. Write Canvas Context to portfolio.json

Add or update the canvas_context object in portfolio.json:

{
  "canvas_context": {
    "canvas_version": "1.0",
    "canvas_date": "2025-01-16",
    "source_file": "current-canvas.md",
    "problems": ["Primary: ...", "Secondary: ..."],
    "uvp": "World's first open-source...",
    "channels": ["Partner Network", "Open Source", "Digital Marketing", "Direct SME Sales"],
    "cost_structure": ["Platform Development", "Partner Program", ...],
    "key_metrics": null,
    "unfair_advantage": null,
    "assumptions": [
      "SMEs will pay €25-40k for partner-led transformation",
      "Consultants will join and actively participate in collective",
      ...
    ],
    "revenue_seeds": {
      "builder-license": { "type": "subscription", "assumption": "€15k/year" },
      "certifications": { "type": "project", "assumption": "€3k/participant" },
      "consulting": { "type": "project", "assumption": "€15-45k/project" },
      "maintenance": { "type": "subscription", "assumption": "€2k/year for 1-5 workplaces" }
    }
  }
}

8. Sync and Summarize

Run the portfolio sync script if products were created:

$CLAUDE_PLUGIN_ROOT/scripts/sync-portfolio.sh <project-dir>

Present a summary:

Entity Type	Created	From Canvas Section
Products	3	Solution (3 layers)
Features	6	Solution (capabilities)
Markets	3	Customer Segments (3 tiers)
Canvas Context	saved	Problem, UVP, Channels, Revenue, Costs, Assumptions

9. Recommend Next Steps

The founding-stage portfolio is now seeded. Suggest the natural next steps, calibrated to the company's stage:

1. Visualize the structure with the portfolio-architecture skill — see the product-feature hierarchy at a glance, spot empty products or unbalanced feature distribution before diving into refinements
2. Refine products with the products skill — challenge positioning, check boundaries between layers
3. Refine features with the features skill — ensure IS-layer descriptions are market-independent
4. Size markets with the markets skill — even rough TAM/SAM/SOM estimates help prioritize
5. Generate propositions with the propositions skill — this is where canvas pain points become DOES/MEANS messaging
6. Define solutions with the solutions skill — canvas pricing assumptions become seed values
7. Validate assumptions — the assumptions from the canvas are preserved in portfolio.json; revisit them as the business evolves

If sections were unfilled (Key Metrics, Unfair Advantage), note that these can inform the portfolio later as the business matures.

Handling Canvas Variants

Business Model Canvas (BMC)

If the user provides a full BMC instead of a lean canvas, the mapping adjusts:

• Key Partners → stored in canvas_context (informs partnership solutions)
• Key Activities → may inform features
• Key Resources → stored in canvas_context
• Value Propositions → same as UVP
• Customer Relationships → stored in canvas_context
• Customer Segments → same mapping as lean canvas
• Channels → same mapping as lean canvas
• Cost Structure → same mapping as lean canvas
• Revenue Streams → same mapping as lean canvas

Incomplete Canvases

Early drafts often have empty sections or "?" placeholders. Handle gracefully:

• Create entities only from sections with actual content
• Note unfilled sections in the summary
• Suggest the user fill gaps before running downstream skills (especially propositions, which need both features AND markets)

Important Notes

• Always confirm entities with the user before writing — never auto-create
• All products get maturity: "concept" unless the user says otherwise
• All features get readiness: "planned" unless the user indicates something is already built
• Markets default to dach region — ask the user to confirm the operating region
• Revenue stream pricing is stored as context, not committed to solution files (the solutions skill handles proper pricing)
• Canvas assumptions are preserved, not discarded — they're the founding team's hypotheses and should stay visible throughout the portfolio lifecycle
• Communication Language: Read portfolio.json in the project root. If a language field is present, communicate with the user in that language. If no language field is present, default to English.
• Refer to $CLAUDE_PLUGIN_ROOT/references/data-model.md for complete entity schemas