PROMPT-LIBRARY.md

# Horizon Forge prompt library

Copy one request into a new task with the plugin installed. Dates are derived from the current date unless you specify them.

## Broad discovery

```text
Use $horizon-forecast to find overlooked industries across the global economy that could grow substantially over the next 5–10 years. Plan first, then execute with specialist agents. Cover physical infrastructure, industrial services, demographic needs, and emerging markets as well as technology. Give me a long, carefully researched dossier with growth ratings, overlooked-opportunity ratings, confidence, scenarios, counterarguments, and signposts. Identify the smaller enabling segments inside larger trends.
```

## A field in depth

```text
Use $horizon-forecast to analyze the future of [FIELD] over 5 and 10 years, with 1–3-year signposts and 10–20-year possibilities. First create the research plan and procedure, then execute them. Study who pays, economics, adoption, supply constraints, institutions, substitutes, regional differences, and second-order effects. Identify which subsectors are likely to capture economic value and which popular claims lack evidence. Produce the complete research dossier.
```

## Compare competing opportunities

```text
Use $horizon-forecast to compare [FIELD A], [FIELD B], and [FIELD C] for growth in real annual economic size over the next decade. Normalize their market boundaries before comparing them. Show absolute additions as well as growth rates. Rank the evidence-backed cases, keep speculative ones separate, and explain the assumptions that would reverse the order.
```

## Look for overlooked enabling businesses

```text
Use $horizon-scan to find neglected suppliers, maintenance businesses, compliance services, workforce needs, and infrastructure that could benefit from [MAJOR TREND]. Follow the causal links from deployment to actual customer spending. Explain why each opportunity may be underprovided, what could commoditize it, and which ones deserve a full forecast. Do not assume low media attention means good economics.
```

## Challenge an exciting thesis

```text
Use $horizon-red-team on this thesis: [THESIS OR REPORT]. Find the strongest external counterevidence, failed analogues, constraints, substitutes, and market-accounting errors. Identify the premise most likely to be wrong, the observation that would invalidate the conclusion, and which parts survive the challenge. Preserve uncertainty instead of manufacturing a definite verdict.
```

## Reconcile conflicting market reports

```text
Use $horizon-evidence to reconcile these market estimates: [LINKS OR FILES]. Trace their original data, definitions, dates, geography, currency, and annual versus cumulative figures. Show which are genuinely comparable, which repeat the same source, and whether a defensible baseline can be established.
```

## Update a saved forecast

```text
Use $horizon-update to revisit [SAVED FORECAST]. Check its registered signposts and new evidence. Keep the original horizon and forecast history intact. Explain every material probability or ranking change, resolve events only under their original rules, and score resolved predictions when a comparable evaluation is possible.
```

## Turn a forecast into business research

```text
Use the findings in [SAVED DOSSIER] to investigate realistic business entry points for [PARTICIPANT TYPE, REGION, AND CONSTRAINTS]. Distinguish industry growth from what this participant can capture. Research the customer, budget, distribution, capital requirements, competition, and smallest useful validation step. Keep the industry ranking separate from the new accessibility assessment.
```

## Focus additional effort where it helps

```text
Take another research pass on the two assumptions most likely to change this ranking. State the question and the evidence that would change the answer, gather independent evidence, then revise the report if warranted. Do not expand the report with tangents that do not affect the decision.
```