# Parent comparison: image 035

Parent: /root, local task 01a0e067-c27c-77f2-a3f9-ec52a44b62cf. Fresh describer: /root/blind_035, gpt-6-luna high, forked with no conversation context. Its only evidence input was this image. Parent read the completed description and verified the image checksum.

[Actual screenshot](../inputs/035.png) · [Independent description](../descriptions/035.md)

Variant B. Captured action: Search Test Garden after manual sent status and saved address. URL: https://kertujitok-proposal-b-20260927.clientsflowdigital.workers.dev/panel/proposals/. Screenshot SHA256: `963955c59a4e62727c99f2380deff0a13796a39d76a1bcf922d6c7829f2f3256`.

- C02: When the Test Garden lead and proposal are edited and saved in variant A, then those changes survive a reload in A while the same starting records in variant B remain unchanged. When the edits are repeated in B and interleaved with A saves, then each variant retains its own distinct saved values and never overwrites the other.
- C04: When Péter searches Ajánlatok for Test Garden and opens its row, then the correct saved proposal opens with its current totals, draft/sent label and document history.

**Parent comparison:** The independent description identifies the actual Test Garden search field, grouped draft/sent results, correct 374,500 Ft saved source card and customer entry points in B. This supports the search/list state. Reopen, saved status and history are evidenced separately by screenshots 036/043 and the completed GUI readbacks. These are manual labels, not evidence of a sent message.

[Completed GUI journeys and exact behavioral evidence](../../deployed-gui/scenario-matrix.md). Full criterion acceptance also requires the other applicable states and nonvisual checks. A screenshot alone does not prove persistence, timing, arithmetic or downloaded bytes.