Capability Map
What can we actually do toward turning skills into businesses?
Capability Assessment
| # | Capability | Maturity (0-4) | Evidence | Risk |
|---|---|---|---|---|
| 1 | Idea entry (venture_business_ideas) | 2 | Table exists in schema, fields defined, ventureId FK ready | LOW — schema work done |
| 2 | AI domain research | 0 | Not built. No LLM pipeline for market/competitor analysis | HIGH — core function missing |
| 3 | Value proposition generation | 0 | Not built. No template, no scoring, no eval gate | HIGH — core function missing |
| 4 | ICP identification | 1 | create-icp skill exists in dream repo. Not in production code. | MEDIUM — pattern exists |
| 5 | VFL scoring (vfl-force-potential) | 1 | Algorithm designed in PRD. No implementation. | HIGH — gate function missing |
| 6 | Eval runner | 1 | Algorithm designed. LLM-as-Judge pattern documented. | HIGH — quality gate missing |
| 7 | SPCL scoring | 1 | Algorithm designed. Content prioritization logic documented. | MEDIUM — Phase 2 dependency |
| 8 | Explore-exploit | 1 | Multi-armed bandit algorithm designed. No implementation. | MEDIUM — Phase 1 dependency |
| 9 | Optimal stopping | 1 | 37% rule algorithm designed. No implementation. | MEDIUM — gate function |
| 10 | Sales forecasting | 1 | Algorithm designed. L0 in CRM dashboard. | LOW — existing foundation |
| 11 | WorkChart generation | 1 | work_chart_definitions table exists, ventureId-scoped | MEDIUM — templates needed |
| 12 | Sprout Generator | 2 | Built and registered, migration applied, blocked at first run | HIGH — prerequisite |
| 13 | Agent assignment | 0 | Activation backlog tables in schema, no implementation | LOW — Phase 3 |
| 14 | Compound rate tracking | 1 | Algorithm designed. No implementation. | LOW — Phase 4 |
Maturity Distribution
| Level | Count | Capabilities |
|---|---|---|
| 0 (Gap) | 3 | AI domain research, value prop generation, agent assignment |
| 1 (Designed) | 8 | ICP, VFL, eval-runner, SPCL, explore-exploit, optimal-stopping, WorkCharts, compound rate |
| 2 (Built) | 2 | Idea entry (schema), Sprout Generator (blocked) |
| 3 (Wired) | 0 | — |
| 4 (Proven) | 0 | — |
Schema Surface (Existing)
The database schema already encodes the idea-to-venture flow:
| Table | Purpose | Maturity | Phase |
|---|---|---|---|
| venture_business_ideas | Entry point: idea in, scored, gated | 2 (exists) | 1 |
| venture_discovery_sessions | Market intelligence capture | 2 (exists) | 1 |
| knowledge_base | Domain research storage | 2 (exists) | 1 |
| venture_contacts | Seed ICP profiles | 2 (exists) | 2 |
| work_chart_definitions | Operations templates (ventureId-scoped) | 2 (exists) | 2 |
| planning_projects | Project scaffolding | 2 (exists) | 2 |
| cost_centers | Accounting structure | 2 (exists) | 2 |
| gl_accounts | Financial tracking | 2 (exists) | 2 |
Gap Analysis
| Gap | Severity | Blocks |
|---|---|---|
| No end-to-end flow tested | Critical | Can't validate any capability in context |
| Zero algorithms implemented | Critical | No scoring, no gating, no optimization |
| No AI pipeline for research/generation | High | Core functions of Phase 1 don't exist |
| No eval quality gates | High | AI output quality unverified |
| No user validation | Medium | Persona and struggling moment unconfirmed |
Key Finding
9 of 14 capabilities sit at maturity 0-1. The schema surface is stronger than the logic layer — tables exist but nothing connects them. The riskiest gap isn't any single capability, it's the absence of a single tested path from idea entry to venture creation. One end-to-end flow would validate (or invalidate) the entire architecture.
Context
- Dependency Map — Previous: what must happen first?
- Agent & Instrument Diagram — Next: how do agents orchestrate?
- Intelligence Functions — Where algorithms get implemented