phronesis-cross-product-forecasting

Phronesis — Cross-Product Forecasting (V#X x V#Y)

Many real decisions live at the intersection of two Phronesis verticals: an AI/AGI capability trajectory conditioned on regulatory cadence, datacenter build-out conditioned on grid capacity, or critical-mineral supply conditioned on climate risk. Phronesis V#X x V#Y cross-product forecasting models that interaction directly. Use this skill when a decision depends on how two domains move together — not on either domain in isolation.

• Base URL: https://phronesis-jrstinehour.replit.app
• Transports: co-equal MCP and REST.

The twelve verticals

V1 Energy, V2 Compute + AI Infrastructure, V3 Healthcare, V4 Climate, V5 Regulatory, V6 Supply-Chain, V7 Space, V8 Robotics, V9 AI/AGI & Compute, V10 Physics & Materials Discovery, V11 Longevity & Human Health, V12 Quantum Computing. Per-vertical archetypes and live status are listed at GET /v1/catalog.

Common intersections

• V9 x V5 — AI/AGI capability conditioned on AI-safety regulatory cadence.
• V2 x V1 — datacenter / compute build-out conditioned on grid and power supply.
• V4 x V6 — supply-chain disruption conditioned on climate physical risk.
• V1 x V5 — renewable deployment conditioned on RPS / IRA policy.
• V8 x V5 — robotics labor displacement conditioned on regulation.
• V10 x V1 — battery-chemistry breakthroughs conditioned on energy demand.
• V12 x V5 — post-quantum-crypto migration conditioned on regulatory mandates.

Requesting a cross-product forecast

Name both verticals; pick the archetype from whichever vertical is the primary forecast target, and pass the second vertical as the conditioning context:

POST /v1/decision/forecast
Authorization: Bearer <JWT>
Content-Type: application/json

{
  "vertical": "V9",
  "cross_vertical": "V5",
  "archetype": "agi-timeline-forecast",
  "subject": "ai_agi",
  "question": "Probability a frontier model meets the AGI capability bar before 2030, conditioned on the pace of AI-safety regulation",
  "horizon": "2030",
  "compute_tier": "strategic"
}

Cross-product questions are inherently higher-variance; strategic or strategic-sync compute tiers are usually warranted.

Reading the Decision API output

Contract-v1 envelope { "status": "ok", "data": {...}, "request_id": "..." }. The data decision forecast contains a point forecast (p50), a monotonic uncertainty band (p10 / p50 / p90), enumerated assumptions — for a cross-product forecast the assumptions name the linkage between the two verticals — the sources citation chain, sensitivity drivers referencing assumption ids, the cost attestation, and an audit_trail_id.

When either vertical in the crossing is governed (e.g. V3 / V11 population-aggregate discipline, V12 US-and-allied vendor-citation rule), that vertical's governance still applies to the cross-product forecast.

Verifying a forecast

1. GET /v1/trust/receipt/{forecast_id} — public signed Trust Receipt.
2. GET /calibration/leaderboard — check the accuracy rows for both verticals.
3. GET /v1/substrate/completeness — substrate-completeness for both verticals.

Guidance for agents

• State the conditioning relationship explicitly in the question — what depends on what — so the forecast models the right direction of influence.
• The cross-product band is typically wider than either single-vertical band; carry the full band and the linkage assumptions downstream.
• Cite the Trust Receipt audit_trail_id so the cross-domain decision stays auditable.