Our UR5e test stopped after placement but before the pocket update. On restart it offers that pocket again. One saved number, two different ideas about where the spacer is. What should the recovery review include?
Write the physical observation beside the saved state for that interruption, then test the boundaries around pickup, placement and saving. Include tray and transfer identity. Where the outcome cannot be established, recovery needs an explicit unresolved case, not an automatic repeat based on the old number.
Controls has put the unresolved case into the draft. We can identify the tray but still lack confirmation for the destination at one boundary. That gap is staying visible before operator practice.
Please post what the review decides there. Also check a second restart while that transfer remains unresolved; a careful first screen won't help if the next startup silently rebuilds it as ordinary work.