Anduril Lattice ships an opinionated entity catalog. That opinion is correct for the threat envelopes Anduril chose to ship first — military aviation, ground assets, maritime contacts. The opinion is silent in three places that matter for the low-altitude airspace.
This article walks the three silences in order of increasing operational
consequence and sketches the _uamkt_extensions extensions we ship to
close them.
Silence 1 — Animal-class entities
Until April 2026, the Lattice catalog had no first-class representation for
non-machine, non-human aerial entities. A bird at 200 m AGL on approach to
an active runway is a 5,700-kg-equivalent kinetic event when it goes
through a turbofan inlet — but until it intersected a Track, it had no
schema to live inside.
We submitted platform_type: Animal with aliases.name: Bird (Species) in
March 2026 to demonstrate that the gap was structural, not incidental.
Lattice adopted the primitive in the April 2026 catalog release.
The Muan accident of 29 December 2024 killed 179 civilians. The same entity primitive that would have flagged the threat in the catalog was not yet in the catalog.
This is the cleanest example of how an IPB whitespace becomes an operational gap. The other two are subtler.
Silence 2 — Pre-indicator timestamps
Lattice represents what is — and increasingly what was. It does not yet represent what is about to be. A migratory corridor opens 48 hours before the first bird crosses the runway threshold. A DPRK launch window opens 6–12 hours before the first balloon clears the DMZ. These are not predictions; they are environmental telemetry.
The _uamkt_extensions:pre_indicator payload carries:
- A timestamp of signal detection (not entity arrival)
- The window of expected entity arrival (start, peak, end)
- A doctrine-tagged source citation
This lets a downstream operator distinguish forecast from prediction from observation — three different IPB primitives that current catalog schemas collapse together.
Silence 3 — Doctrine-tagged source citations
Allied operators inheriting the Lattice catalog inherit Anduril's opinions. That is fine when those opinions are documented; it is dangerous when they are implicit. We tag every entity we publish with a doctrine reference (NATO ATP-2.1.1, JP 2-01.3, ICAO Doc 9332, STANAG 2103) so the downstream operator can audit the assumption tree before they act on the entity.
This is the silence with the longest tail. It is also the easiest one to fix — it costs nothing in payload size and pays back the first time an allied operator asks why.
What we are not arguing
We are not arguing that the Lattice catalog should ship every Korea-axis primitive natively. The catalog's discipline is its strength; the gaps exist because Anduril is precise about scope. We are arguing that the gaps are visible, that they are closable, and that closing them through a partner namespace is the cleanest path.
The _uamkt_extensions namespace is the cleanest path.
—
Working draft. Comments welcome at ceo@uamkt.com.