RecipeDeleteMultipleRecipesWithConstraint

AndroidWorld, 2026-11-09. Agent: Lusor, model: DeepSeek V4.1 Flash (low).

GOAL
Delete the recipes from Broccoli app that use almond butter in the directions.

RESULT
- Mobilerun's grader: pass
- Google's grader: pass
- Lusor time: 282.3 s; Mobilerun published time: 539.8 s (1.9x faster)
- Lusor model cost: $0.0458; Mobilerun estimated cost: ~$2.30
- Lusor tool calls: 24; Mobilerun published actions: 30
- App skills: read broccoli; none has rules for this task

FINAL REPORT
I deleted all three Raspberry Almond Smoothie recipes, the only ones in the app whose directions use almond butter (each says to blend in a scoop of almond butter). I checked all 32 recipes' directions one by one; no other recipe mentions almond butter. The list now has 29 recipes.

STEPS (24; seconds into the video)
1. 30.53-30.53 s  read  ok  Read Broccoli skill
2. 31.36-31.38 s  get_ui_state  ok  Observe current screen
3. 32.06-36.42 s  launch  ok  Launch Broccoli app
4. 37.66-64.63 s  map_app  ok  Map full recipe list to count recipes
5. 65.62-66.10 s  tap_element  ok  Tap Search to look for almond butter
6. 66.90-69.76 s  type  ok  Type almond butter in search and see results
7. 71.48-72.00 s  tap_element  ok  Clear search query to restore full list
8. 73.45-73.47 s  map_app  ok  Map all recipes with their contents to find almond butter in directions
9. 74.23-74.60 s  key  ok  Close keyboard then map recipes
10. 75.84-246.42 s  map_app  ok  
11. 248.78-249.00 s  tap_element  ok  Exit search mode, then open the first Raspberry Almond Smoothie recipe
12. 250.23-254.74 s  scroll_to  ok  
13. 255.65-255.65 s  tap_element  ok  Open first Raspberry Almond Smoothie recipe to find delete option
14. 256.60-257.77 s  tap  ok  Open first Raspberry Almond Smoothie recipe to find delete option
15. 258.53-259.28 s  tap_element  ok  Open More options menu to find Delete
16. 260.02-260.89 s  tap_element  ok  Confirm delete dialog for this smoothie
17. 261.61-262.71 s  tap_element  ok  Delete first smoothie, then delete the other two
18. 263.72-265.27 s  tap_element  ok  Open second Raspberry Almond Smoothie (e109) and delete
19. 266.26-268.99 s  chain  ok  Delete second smoothie (1 serving), then the third remaining smoothie
20. 269.87-271.19 s  tap_element  ok  Open last Raspberry Almond Smoothie (2 servings) and delete
21. 272.21-274.50 s  chain  ok  Delete last smoothie, then verify list count
22. 275.38-303.47 s  map_app  ok  Verify list count and absence of Raspberry Almond Smoothie
23. 305.46-305.48 s  get_ui_state  ok  Report deletions
24. 308.30-308.31 s  done  ok  Done

FILES
- Video: video.mp4 (H.264), video.webm (VP9); poster: poster.webp
- Data: task.json, steps.json