# First and Last > Re-order a short list so the things that must be remembered sit first and last. Write what the list is for > (a doorstep handover, a sign, steps read over the phone, the points of a short talk) and three to ten items, > one per line. The page pegs them on a washing line that sags in the middle - the dip where list items are > easily lost - marks the strong spots at each end, and lets you star, drag, nudge with the arrow keys or > auto-place starred items. A metered run judges which lines matter most for the goal, whether the order can > change, and either moves what matters to the ends with a reason per move, or keeps fixed steps in order with > an opening line and a recap. Built on the SkillSafe skill @owl-listener/serial-position-effect. https://first-and-last.skillsafe.ai/ ## What is free - The washing line, the strong-spot markings and an illustrative stick bar per place (the usual serial-position shape, not a measurement). - Starring, dragging, arrow-key and button re-ordering, and "Move starred to the ends" (a fixed placement rule). - Three sample lists replay their saved placements at no cost. - Downloads: the line (PNG, SVG), the list in order (plain text, Markdown, JSON), and a Share button where the browser has a share sheet. ## The metered task One task, `order`, on gpt-terra, metered per run. The input is string fields: `task`, `goal`, `items` (numbered lines `N. the item`), `item_count`, `starred` (line numbers or `none`), `order_hint` (`fixed`, `free` or `unsure`), `strong_spots`, `fill_order` and `pick_count`. The reply is one JSON object: `verdict` (`bookend` - the order must stay; `keep` - what matters already sits at the ends; `reorder`), `key` (the lines that matter most, most important first), `order`, `moves` (a reason per moved line), `drop` (at most two lines the goal does not need), `opener`, `recap` and `note`. Strong spots: the first and last place for three or four items, the first two and the last place for five, the first two and last two for six or more. Key lines fill them first place, last place, second, second-to-last; everything else keeps its own order. The page recomputes that placement itself and checks every reply against it. ## Links - API tutorial: https://first-and-last.skillsafe.ai/api.html - Source skill: https://skillsafe.ai/skill/@owl-listener/serial-position-effect (owl-listener/designer-skills)