Dialogs of kept parts open only in edit mode in a visible place; ids
rule for both annotation bodies; kept parts mount at first edit render;
repair geometries lifted; N2, N4, N6, N7 stated.
Whole shell suite at B1, E, F; unmounted-forms clarification prevails
over the plan lines, B2 and H lists; distinct ids for read-only twins;
drawing actions kept hidden with dialogs closed; 9 MINOR answered.
AC3, AC4, AC5, AC7, AC17 and changes 2.6-2.9 applied as approved
(BDR-105); per-operation sandbox go; precautions 2.8 (no silent loss
on a 1 024 px crossing or recovery decision) and AC7 (save time rights)
resolved from the code facts, AC21 and AC22 added. Plan unchanged.
Owner brief « Ratification finale U1 et autorisation conditionnelle
d'implémentation » (BDR-105): study §6 tests, §7 table, 7.2, 7.3,
UX plan U2b entry, REVUE and decisions pointer. No implementation.
Owner arbitrations: F2 leaves, F3 fields, F4 model A, F8, F22 in the study
with their tests. U1 checkpoint page: 18 choices, exact AC amendments,
audits traced, sandbox reset scope, criteria per lot; no implicit deletion.
Owner principle validation recorded: U2b after U2; full decision table with
alternatives, consequences and state; F2 groups, F3 field list and F4 view
semantics computed from the equine draft; boards shown with their decisions.
Verifier round 2: view chip and grasset names computed from the draft;
partial clears, out-of-filter count, zone picker order, dashed sibling
and result presentation listed as owner decisions.
Owner brief: bidirectional navigation between annotations and findings.
Relations audit (one finding_zone_id link, no new association), six
boards from the equine draft data, placement U2b, risks, tests, owner
decisions F1-F16. Design only; U1 plan untouched.
Owner approach A and four choices recorded as gated clarifications.
Plan lots A to I after three blind challengers and one confirmation
pass; no application code, no DEV.