Test Scenario: Verify day-scaled kappa/sigma calibration fix for IR curve feeds

This page documents a test scenario verifying Fix day-scaled kappa/sigma calibration across short-rate processes in IR Rates synthetic data: dataset seeding, index cleanup, dual-curve, quoting conventions. It is filled in with the target dialog and checklist of steps before testing starts; the QA Validation Runner panel rewrites * Results in place on save.

Before the fix, starting an IR curve config published wildly unrealistic rates (~150-250%, and a FRA point that never moved) because discount_factor() accumulated bond-time as if every tick were a full year regardless of what a tick represents. This scenario starts a real, provisioned config (synthetic.ir_curve_configs.realistic, CIR-based — this is the dataset the Barclays provisioning flow actually publishes; Vasicek's own discount_factor()=/=next() dt-scaling is already covered by 166 automated Catch2 test cases including dedicated per-process dt coverage, so this scenario's job is an end-to-end UI-level sanity check of the published numbers, not a re-verification of the per-process math) and confirms the published rate is a plausible single-digit percentage that actually moves across tenors, not a blown-up or frozen one.