the reading item mint has no caller outside its own package so the admission and surprise schemas gate a corpus nothing can populate

the reading item mint has no caller outside its own package so the admission and surprise schemas gate a corpus nothing can populate

Found by the itd-189 fidelity audit and verified independently by the orchestrator. This is a "wired or it isn't done" finding.

capture.IngestReading mints rdi-N. Grepping the whole tree for callers finds only its own test file: no CLI verb, no plugin surface, nothing under cmd/. The CLI has verbs that CONSUME a reading item -- capture promote rdi-N and capture disposition rdi-N -- and nothing that creates one.

So itd-189's admission and surprise schemas, and the outstanding-readings report that reads them, are armed over a corpus the shipped product cannot populate. Every acceptance criterion's Given clause presupposes a reading item that no shipped surface can produce.

The producer is itd-185, the ingest door, which is planned/ and unstarted. Under itd-192 that makes the criteria MET_WITH_CONCERNS rather than NOT_MET, because this phase does wire it -- and the auditor named itd-185 correctly.

This materially widens iss-2608301726130926, the armed Phase-4 conditional. That record currently says two criteria flip to NOT_MET if itd-185 does not land: itd-178 ac-2 and itd-180 ac-1. Itd-189's three criteria now rest on the same intent. The conditional's blast radius is at least five criteria across three intents, not two across two.

Grounds

• pursued: the record writer gains its intended caller and the reading-record family becomes populatable from the shipped product; a run that ingested and left no rdi-N record would show this wrong