Trend Research

Research groundwork for the TIPS trend pipeline. Reads agreed trend-scout candidates, optionally deep-researches 3–5 high-value Act-horizon trends, dispatches 4 parallel agents (one per Smarter Service dimension) to enrich every candidate with web-sourced quantitative evidence and extract claims, then writes a single research manifest that downstream synthesis and booklet skills consume as their gate.

Purpose

Produce the canonical research artefacts that both the TIPS report (/trend-synthesis) and the comprehensive TIPS catalog (/trend-booklet) need:

1. Validate prerequisites (trend-scout agreed, value-model with investment themes)
2. Select 3–5 high-value Act-horizon trends for optional deep research
3. Enrich each of the ~60 candidates with quantitative evidence and source citations via 4 parallel cogni-trends:trend-report-writer agents
4. Extract per-trend claims for downstream verification
5. Write .metadata/trend-research-output.json — the single manifest both /trend-synthesis and /trend-booklet gate on
6. Auto-recommend /trend-synthesis (canonical TIPS report) and /trend-booklet (full candidate catalog)

This skill does not write the final report — prose composition lives in /trend-synthesis. Splitting research from synthesis lets the user re-run either stage independently and lets the booklet skill consume the same evidence pool without duplicating work.

Language Support