skills/horizon-forecast/templates/research-plan.md

# Research plan: [topic]

## Decision and scope

- User's decision and intended use:
- As-of date and information cutoff:
- Geography and regional coverage:
- Baseline metric, currency/price basis, and market boundaries:
- Near-term, primary, and longer-term endpoints:
- Definition of substantial expansion and resolution source:
- Depth/deliverables and any user constraints:
- Assumptions being used while optional clarifications are pending:

## Ordered procedure

| Stage | Research question | Method/source targets | Owner | Dependencies | Completion evidence | Status |
|---|---|---|---|---|---|---|
| Frame | What exactly are we comparing? | Definitions and reference classes | Director | None | Comparable boundaries and event definition | Pending |
| Discover | Which segments deserve deeper work? | Diverse economic-driver scan | Scout/director | Frame | Candidate map and exclusions | Pending |
| Establish | What is observed today? | Original datasets and demand/cost records | Specialists | Candidate map | Claim/source ledger | Pending |
| Connect | What enables or prevents growth? | Causal/dependency map and regional constraints | Specialists/director | Initial evidence | Material tangents and critical path | Pending |
| Quantify | What outcomes fit demand and capacity? | Bottom-up/top-down checks and scenarios | Forecaster/economist | Compatible baseline | Reproducible inputs, outcomes, sensitivities | Pending |
| Rate | How do opportunities compare? | Anchored rubrics and confidence | Director | Quantification | Explained scorecards and tiers | Pending |
| Challenge | What could reverse the conclusion? | External counterevidence and numerical audit | Independent reviewers | Initial draft | Findings and dispositions | Pending |
| Resolve | Which gaps need another pass? | Targeted answer-changing research | Director/specialists | Challenge | Revised results or explicit limitations | Pending |
| Deliver | Can the user assess and revisit the forecast? | Final checks and forecast register | Director | Material findings handled | Dossier and update signposts | Pending |

## Agent and evidence protocol

List actual available slots, assignments, independent initial estimates, merge ownership, and fallback if tools fail. Specify authoritative source types, critical corroboration needs, geography gaps, and prohibited/inaccessible data.

## Ranking choices and depth allocation

Freeze weights, comparison-specific scale bands, shortlist targets, thresholds, and high-value research questions. Define tangent relevance and the stopping convention. Do not change the rubric after seeing results without recording why.

## Completion and departures

Update with completed, omitted, and blocked stages; explain material departures from this plan and implications for the final answer.