Train-related prompt scripts surge; small declines across ESX modules

Sustained adoption is the story. The week saw 1,601,275 distinct scripts observed across servers (previous week 1,581,175), but meaningful movement clustered in a handful of names. The clearest winners were five prompt_train* resources that all recorded growth on every measured day; each moved from the mid‑20s to the mid‑200s of server appearances, which is a rare magnitude of spread over a single week. Outside that family, most large, familiar packages were flat or down by a couple of percent — the platform’s core modules remain widely deployed, but small, steady churn still nudged their counts downward. New releases appeared (see newcomers) and one previously visible mod vanished completely (see declining). This is not a week of broad platform reshaping; it is a week in which a single cluster found traction while incumbent libraries saw modest erosion. For developers and server owners the salient signal is the shape of movement: steady, multi‑day increases imply diffusion across operators; single‑day jumps usually point at packaging or one server’s update. Read the week’s winners and losers with that distinction in mind.

All five train-related prompt_* scripts grew on 6/6 days, and their daily sequences show steady, repeated gains rather than one-off jumps — the kind of adoption pattern that matters. prompt_trainstation_paleto reached 265 servers (previous week 24), adding 241 servers and recording daily counts of 183→217→257→275→295→310→319. prompt_sandy_traintunnel reached 262 servers (previous week 24), up 238 servers with a daily run of 181→215→256→269→290→307→313. prompt_trainstation_ls sits at 256 servers (previous week 23), having added 232 servers and moving 177→207→251→266→284→297→306 across the week. prompt_train_polar_express reached 249 servers (previous week 24), a +226 increase with daily 178→208→246→255→272→288→297. prompt_train_script rose to 226 servers (previous week 22), adding 204 servers and following 170→191→225→235→247→254→259. Those five appear as a coherent cluster: they share a name prefix, they all grew every day, and each gained a few dozen servers on top of an already visible base. That combination — common naming plus matched temporal shape — can indicate a single author, a bundled feature set, or a piece of content that multiple servers are installing in the same way; the data alone does not prove which. What the numbers do show is diffusion: daily increases across six days are what you look for when deciding whether a resource is spreading by recommendation or compatibility rather than being adopted only where a single operator pushed an update. For developers, this is a watch‑list: code maintained in this family is currently finding installations quickly enough to merit compatibility checks and dependency reviews. For server owners, the steady growth suggests these resources are being integrated into a range of server builds rather than riding a single large server’s release.

A cluster of long-standing ESX components lost small amounts of reach this week. es_extended fell from 12,397 to 12,128 servers (-269, -2.2%) and grew on only 2/6 days; its daily counts read 12,197→12,137→12,106→12,123→12,107→12,118→12,109, a shallow downward slope with brief recoveries. The esx_menu family — esx_menu_list (10,539→10,539? see note) — recorded a net loss (esx_menu_list at 10,539 servers, down 256), esx_menu_dialog (10,488, -253) and esx_menu_default (10,576, -252) all declined in the same 2–2.5% band and similarly showed intermittent upticks over the seven daily measures. Other ESX staples followed the same pattern: esx_status lost 253 servers to 10,057 (-2.5%), esx_addonaccount fell by 245 to 10,197, esx_society declined to 9,566 and esx_datastore and esx_basicneeds each shed around 240 servers. Those moves are large in absolute terms but small proportionally, and the day‑by‑day sequences show slow attrition rather than a sudden failure or mass uninstall. PolyZone lost 228 servers to 17,633 (-1.3%) while still appearing on more than 17,600 servers; its daily counts (17,720→17,637→17,568→17,584→17,581→17,647→17,695) mirror the same gentle wobble. Separately, ; its last sighting was 2026-06-25. A disappearance at that scale usually indicates a removed package, a broken update, or deprecation of a private mod; the data cannot distinguish which, but the abrupt absence is worth noting for anyone depending on it.

This week produced a modest crop of first‑seen names. ak47_fpsbooster is the largest of the newcomers, first observed 2026-06-22 and reaching 106 servers in its first week — a notable debut because it cleared triple digits immediately. d7cyp (first seen 2026-06-23) reached 94 servers, another rare high‑double‑digit entry for a first week. A handful of smaller new names arrived: custom_store_mrpd reached 30 servers, cs-sellped 23, krezma_tech 18, and a set of OL-* UI pieces (OL-hud 18, OL-inventory 17, OL-menu 17, OL-ui 16) each reached the mid‑teens; motelonwheels also reached 16 servers after first appearing 2026-06-22. Most new resource names are private or single‑server projects; the ones to watch are those that clear the dozens or hundreds mark in their first week because they are uncommon and usually indicate a public release, a bundled package, or a rapid pickup by template servers. ak47_fpsbooster and d7cyp fall into that category this week.

What to watch next week: the prompt_train* family is the clearest candidate for continued growth — their matched, six‑day climb suggests diffusion rather than a one‑off bulk install, and a repeat of this shape would confirm they’ve found broad utility. Developers of other train or transport content should check compatibility; server owners using train-themed resources should test these in staging to see if they solve missing features. For ESX maintainers the declines are a small but visible reminder that even established modules lose reach over time; the pattern here looks like slow attrition, not a single catastrophic event, so keep monitoring for continued erosion or a reversal. Newcomer packs that hit dozens of servers in week one (ak47_fpsbooster, d7cyp) deserve a brief review before depending on them — first‑week traction is meaningful but not definitive. Finally, treat disappearances (xN93_Boat_Yacht) as a red flag for dependencies you don’t control: if something you rely on vanishes from public scanning, consider finding maintained alternatives or vendoring critical code. The central lesson remains the same: adoption is slow and shape matters; multiple days of upward movement are the only reliable signal that a resource is actually spreading.