MAREF PMM Research Skills — Self-Assessment Demo
(Eating our own dog food: validating MAREF's positioning with our own Skills)

======================================================================
STUDY 1: Positioning Validation (7-question framework)
======================================================================
mode: self_assessment
note: Self-assessment mode: no persona panel recruited. The scorecard below reflects a structured gap analysis using the product's own positioning artifacts, NOT market validation. A real validation requires a recruited panel (e.g., Ditto API or human study).

Study design (7 questions):
  Q1 [Competitive Alternatives]: When you think about agent governance for production deployments, what's the fir...
  Q2 [Status Quo + Gaps]: Walk me through how you currently solve agent governance for production deployme...
  Q3 [Value Resonance]: If I told you there was a product that TLA+ formal verification with 5 proven th...
  Q4 [Market Category]: How would you describe MAREF to a colleague? What category would you put it in?...
  Q5 [Competitive Differentiation]: Compared to LangGraph, CrewAI, AutoGen, what would make you choose a new option?...
  Q6 [Primary Value Driver]: If MAREF could only do ONE thing brilliantly for you, what should that be? Why d...
  Q7 [Adoption Barriers]: What would stop you from trying something like this? What would you need to see ...

Positioning scorecard (1-5, self-assessment):
  competitive_alternatives: 5/5 — Positioning names 3 competitors: ['LangGraph', 'CrewAI', 'AutoGen']
  value_resonance: 5/5 — Value prop uses concrete verbs (verify, audit, govern) 
  market_category: 5/5 — Description uses category word: yes
  competitive_differentiation: 5/5 — Strong differentiation signals (formal verification)
  primary_value_driver: 4/5 — Value driver inferred from unique_value_prop; real study would reveal which driver resonates most.
  adoption_barriers: 3/5 — Adoption barriers not yet identified — requires real panel study to surface.

Risk flags:
  ⚠️  Adoption barriers unknown — self-assessment cannot surface them. Requires a recruited persona panel.

Competitive alternative map:
  LangGraph: named_in_positioning=True
  CrewAI: named_in_positioning=True
  AutoGen: named_in_positioning=True

======================================================================
STUDY 2: Messaging Testing (3 tagline variants)
======================================================================
mode: self_assessment
note: Self-assessment mode: no persona panel. Ranking below is a heuristic assessment based on message structure, not market response. A real study requires a recruited panel.

Message performance ranking:
  #1 (variant=problem_led, score=3): 88% of companies had an AI agent incident last year. MAREF i...
  #2 (variant=outcome_led, score=3): Build with LangGraph. Govern with MAREF. Ship to production ...
  #3 (variant=capability_led, score=2): TLA+ verified. 10-state Gray Code. Three-gate skill marketpl...

Recommended primary message: a

======================================================================
STUDY 3: Competitive Intelligence (MAREF vs 3 competitors)
======================================================================
mode: self_assessment
note: Self-assessment mode: competitive perception is inferred from public competitor documentation, not from a recruited panel. A real study would surface perception gaps and landmine questions that this mode cannot.

Competitive perception matrix:
  LangGraph:
    strength: graph-based orchestration, large ecosystem
    weakness: no governance layer, no formal verification
    maref_advantage: MAREF's entire purpose is governance; competitor bolted on none.
  CrewAI:
    strength: role-based agent design, easy to start
    weakness: no runtime safety gates, no audit trail
    maref_advantage: MAREF's entire purpose is governance; competitor bolted on none.
  AutoGen:
    strength: Microsoft-backed, multi-agent conversation
    weakness: no circuit breakers, no skill marketplace
    maref_advantage: MAREF's CircuitBreaker contains failures; competitor has none.

Landmine questions (sales must prepare for):
  💣 If LangGraph adds governance, why do I need MAREF?
  💣 If CrewAI adds governance, why do I need MAREF?
  💣 If AutoGen adds governance, why do I need MAREF?
  💣 TLA+ sounds academic — can you show me a production incident it would have prevented?
  💣 We're already on LangGraph/CrewAI — do we have to rip it out?

Win themes:
  ✅ Production governance gap: LangGraph/CrewAI/AutoGen have 0/10 OWASP coverage.
  ✅ Formal verification: TLA+ proofs are unmatched; competitors have none.
  ✅ Skill marketplace: three-gate admission is a supply-chain differentiator.

Loss themes:
  ❌ Migration anxiety: 'do we have to rip out our existing stack?'
  ❌ Academic perception: TLA+ may feel theoretical to practitioner buyers.
  ❌ Ecosystem size: LangGraph has more integrations and community.

Battlecard (JSON):
{
  "product": "MAREF",
  "positioning": "MAREF is the missing governance layer \u2014 use LangGraph to build, use MAREF to govern.",
  "vs": {
    "LangGraph": {
      "their_strength": "graph-based orchestration, large ecosystem",
      "their_weakness": "no governance layer, no formal verification",
      "maref_wedge": "MAREF's entire purpose is governance; competitor bolted on none.",
      "landmine": "If LangGraph adds governance, why do I need MAREF?"
    },
    "CrewAI": {
      "their_strength": "role-based agent design, easy to start",
      "their_weakness": "no runtime safety gates, no audit trail",
      "maref_wedge": "MAREF's entire purpose is governance; competitor bolted on none.",
      "landmine": "If CrewAI adds governance, why do I need MAREF?"
    },
    "AutoGen": {
      "their_strength": "Microsoft-backed, multi-agent conversation",
      "their_weakness": "no circuit breakers, no skill marketplace",
      "maref_wedge": "MAREF's CircuitBreaker contains failures; competitor has none.",
      "landmine": "If AutoGen adds governance, why do I need MAREF?"
    }
  }
}

======================================================================
SUMMARY
======================================================================
All 3 PMM studies ran in self-assessment mode against MAREF.

What this proves:
  ✅ The 7-question frameworks are correctly encoded and runnable.
  ✅ The Skills produce structured deliverables (scorecard, ranking, battlecard).
  ✅ Self-assessment mode surfaces positioning gaps and risk flags.

What this does NOT prove:
  ❌ Market validation — requires a recruited persona panel (Ditto API or human study).
  ❌ Real competitive perception — inferred from public docs, not panel responses.
  ❌ Adoption barriers — self-assessment cannot surface these.

Next step: acquire Ditto API key and re-run in panel_study mode.
