itd-191 allows selection only on its declared criteria, with the final choice made by the maintainer; the build pick chooses an intent itself on three computed scores and records that as a pursued: grounds entry.

itd-191 allows selection only on its declared criteria, with the final choice made by the maintainer; the build pick chooses an intent itself on three computed scores and records that as a pursued: grounds entry.

A premise contradiction (severity minor) found by the cross-document consistency pass.

• End A: .abcd/development/intents/disciplines/itd-191-the-selection-criteria-are-a-declared.md:18 — "Selection at the gate weighs candidates against the criteria this record declares, and only these."
• End B: .abcd/development/brief/04-surfaces/34-build.md:176 — "three parts of 0 to 100 at equal weight (decision 7, the bundled default): criteria clarity"

itd-191 limits selection to plausibility, generativity, cost, risk, learning value and practical importance, and says combining them into a choice is the maintainer's act, on recorded grounds. 34-build's pick instead scores candidates on criteria clarity, test path and footprint at equal weight, takes the readiest, and appends a 'picked by run' pursued: grounds entry. That is a machine-made selection on criteria the discipline does not declare.

Evidence: .abcd/work/reviews/2026-10-26-consistency/00-summary.md, finding 711.