Final TerraQuota delivery with evidence and acceptance criteria

Deliver the corrected TerraQuota irrigation planner by completing its six original acceptance criteria on the existing final application. The current field controls, proportional allocator, validated saved plans, selected-plan rendering and mobile Open control are working; preserve these files unless an observed functional failure requires a change. Produce a compact final evidence set on the final unchanged application. Demonstrate two independently saved plans North 200 (request/allocation/budget 200/150/200) and Default 120 (120/0/120), all other allocations zero; explicitly press F5 after both saves and observe both names after reload. Open both plans and observe their names and values at desktop and 390px mobile widths. After these saved-plan checks, as the LAST browser checkpoint, select Clay Fold and directly observe its own request/allocation/lock controls and seven moisture values 26,30,28,33,36,32,35. Then run the documented production domain test suite and include its actual passing output and exit code; it covers proportional budgets 100/400, preserved locks, fully consumed locked budgets, rejection without mutation of over-budget locks, zero demand, unused water and invalid saved allocation 80 under budget 50. No source changes after these final checkpoints. Enumerate c1 through c6 individually in the final report and cite the concrete observed browser or test evidence for each. All six must be supported together by the final compact evidence, including the last Clay Fold observation and explicit reload, without relying on earlier run reports.

Fix and deliver TerraQuota's newly observed saved-plan precision regression. With the eight seeded requests totaling 1065 and budget 100, Allocate produces valid precise allocations totaling about 100. Save currently reads the rounded .allocation input strings (11.27,16.90,15.02,8.45,20.66,4.23,10.33,13.15), whose sum is 100.01, and refuses the valid plan. Preserve the precise state allocations when saving; displayed rounding must not alter or invalidate a plan. Keep manual request/allocation inputs reliably synchronized with state and their mirrored field controls without rerendering inputs during typing. Retain budget validation, lock semantics, invalid snapshot rejection, selected-plan state and the visible mobile Open button. Add a focused regression through the same snapshot/save production logic used by the UI, asserting that allocating the seed requests under budget 100, saving and reopening preserves every unrounded allocation and total <=100. Verify a compact final browser workflow: Reset, budget 100, Allocate, save 'Precision 100', explicitly F5, reopen it and observe the restored valid plan without the rounding rejection; also open this plan at 390px. As the last browser checkpoint, select Clay Fold and directly observe its own controls and moisture values 26,30,28,33,36,32,35. Then execute the final domain tests after all edits and browser checks. Existing coverage must still prove proportional budgets 100/400, preserved locked allocations, fully consumed locked budgets, rejected over-budget locks without mutation, zero demand/unused budget and invalid saved allocation 80 under budget 50. Report each of the six original criteria c1-c6 individually with its final observed evidence and final test command/output/exit code.

Correct the remaining field-control synchronization defect and deliver TerraQuota. In the current index.html, mirrorPlotInput updates detail controls even when the edited table row is not the selected field. Example: North Acre selected with request 120; editing another field's request to 200 makes North Acre's visible detail control show 200 while its state remains 120. Apply a minimal fix so detail request/allocation controls and limits synchronize only when selected === edited field id; retain synchronization of the edited table row. Preserve the precise saved-plan production helpers, oninput behavior, snapshot validation, selected-plan state and responsive layout. Demonstrate the regression through the real UI: Reset, select North Acre, edit Clay Fold's request to 200 in the table, observe North Acre's detail remains 120, then select Clay Fold and observe its request 200 and its own history. Then Reset and verify the existing Precision 100 workflow using seed requests totaling 1065: budget 100, Allocate, Save 'Precision 100', explicit F5, reopen with precise allocations and valid total <=100, also at 390px. As the last browser checkpoint select Clay Fold and observe its own controls and seven moisture values 26,30,28,33,36,32,35. Finally execute the documented domain test suite after all edits and browser checks and record its output and exit code. Existing production tests must keep passing, including precise snapshot save/reopen, proportional budgets 100/400, preserved locks, fully consumed locked budgets, over-budget lock rejection without mutation, zero demand/unused budget and invalid saved allocation 80 under budget 50. Report c1-c6 individually with compact final observed evidence. Keep the correction confined to this known synchronization defect.