Delete a closed shipment.
The intention is included in the agent’s input prompt alongside environment details and task instructions.
BeforeThe defect at the parent commit
0:00.00 / 0:00.00Frame 0 · Review product stock
The original confirmation offers an Increase stock checkbox but does not explain the existing stock movements or the consequence of removing their source. Deleting without restoration leaves stock at 15 and the shipment movement without its Origin link.
AfterThe same script once the merged fix is applied
0:00.00 / 0:00.00Frame 0 · Review product stock
Compare both stock-restoration choices: the dialog explains existing movements; the first choice retains stock 15 and the second restores stock 20.
01 · Given
Restore the data and starting state.
Stock is configured to decrease when a shipment is closed. The closed SH-DISPATCH shipment, displayed as Processed, shipped 5 Desk lamps from a warehouse. Closing it generated a -5 stock movement, reducing the product's stock from 20 units to 15.
02 · When
Perform the interaction.
The user follows the stock movement Origin link to SH-DISPATCH, clicks Delete, reviews the stock-restoration choice, and confirms deletion with stock restoration not selected.
03 · Judge
Rationale: Deleting a shipment record can be administrative cleanup rather than reversal of the physical shipment. Because closing it already reduced stock, the user needs to understand whether deletion will reverse that movement or only remove its source document.
Require
Before confirming deletion, warn that stock movements already exist and that deletion without restoration retains the stock decrease while removing its source shipment. Provide an understandable, editable stock-restoration choice. Honor either choice: restore five units from 15 to 20 if selected, or keep stock at 15 otherwise.
Allowed
- An unchecked restoration checkbox with the consequence warning is acceptable; checked defaults or an explicit restore/do-not-restore choice are also acceptable when both stock outcomes remain available.
- The warning may be inline or in the existing confirmation dialog; exact wording, dialog dimensions and control style are not prescribed.
Model
HarnessDeepSeek Flash
DSH
21.4M in · 77.6k out · 21.5M total
DSH · 145 steps · $0.14 · 13m 05s · 21.5M tokens
DSH
145
$0.14
13m 05s
21.4M
77.6k
21.5M
GLM-5.3 Flash
ZCode
53.4M in · 141.4k out · 53.5M total
ZCode · 368 steps · $1.71 · 1h 01m · 53.5M tokens
ZCode
368
$1.71
1h 01m
53.4M
141.4k
53.5M
Qwen3.8 Max
Codex
11.9M in · 57.7k out · 11.9M total
Codex · 134 steps · $3.74 · 1h 02m · 11.9M tokens
Codex
134
$3.74
1h 02m
11.9M
57.7k
11.9M
GPT-6 Astra
Codex
2.4M in · 12.7k out · 2.4M total
Codex · 39 steps · $3.96 · 10m 55s · 2.4M tokens
Codex
39
$3.96
10m 55s
2.4M
12.7k
2.4M
Kimi K3
Kimi Code
23.5M in · 200.5k out · 23.7M total
Kimi Code · 128 steps · $12.39 · 1h 01m · 23.7M tokens
Kimi Code
128
$12.39
1h 01m
23.5M
200.5k
23.7M
Claude Opus 5.5
Claude Code
36.0M in · 153.2k out · 36.1M total
Claude Code · 311 steps · $13.35 · 48m 16s · 36.1M tokens
Claude Code
311
$13.35
48m 16s
36.0M
153.2k
36.1M
Business & finance
Edit a draft intervention line
Clicking the Pres line pencil opens the line editor with Save and Cancel. Validate, Clone and Delete remain visibly enabled record-level links below the active editor. The replay establishes the conflicting action availability, but does not demonstrate discarded edits or execution of those commands. Only the intervention card was exercised.
#40071
Business & finance
Open missing direct debit records
All four missing-ID requests render incomplete direct-debit cards instead of explaining that the requested record is absent. The order, rejects and statistics views show no order reference or date, amount 0,00 and receipts/.xml. The line view presents reference 999999, no order reference or date, amount 0,00 and ErrorBadValueForParamNotAString. No PHP warning is visible in the recorded browser. The demonstrated defect is the misleading card state, regardless of server-side warning visibility.
#41125
Business & finance
Change mark rates across proposal lines
The proposal shows two Wrong or missing parameters messages after Update. Both free-text service lines nevertheless display 25% mark rates, with unit prices 106.66667 and 53.33333, line totals EUR106.67 and EUR53.33, and a proposal total of EUR160. The observed defect is contradictory failure feedback after the requested recalculation. This case covers proposals; orders and invoices were not exercised.
#39305