Include or exclude product types, constrain report dates, drop keyword noise, and filter by threat-language (fear) score against the corpus gauge.
Matches titles and body text. Leave blank to browse by filters only.
Leave all unchecked = search everything. Check one or more to limit results.
Drop types you never want in this result set.
Narrows further when “Side-project series” is included (GhostNet, ATAK, …).
Major areas only. Matches works already tagged with that place/region in the taxonomy (e.g. Middle East, Ukraine, Strait of Hormuz) — not a free-text place search.
Uses report dates on S2 works. Manuals ignore date filters. For a single day, use the calendar.
Drop results whose title/snippet contains any of these words (space or comma separated).
Same units as the home freakout gauge (per 1k words: NRC fear, anger and disgust, minus positive, plus a small S2 threat-lexicon term). Corpus mean 40.26, spread ±19.47, elevated 59.73, crisis 69.46. See metrics for the full distribution and normals by month. Works without a scored body are dropped when this filter is on. Manuals never match.
url_tracker.db