The Dungeon Under the House
A deep dive into Shroud of the Avatar's player-built dungeons, what twenty-five years of player-made adventures taught the genre, and twenty-five concrete proposals.
Any lot owner in New Britannia can hang a dungeon under their house: craft or find the pieces, click a doorway, and assemble halls, crypts, cisterns and boss chambers room by room, then flip the whole thing into Adventure Mode and fight what lives there. That system is deeper than most players realize, and almost nobody visits it. This document tours what shipped, surveys every serious attempt at player-made dungeons the genre has produced, and proposes twenty-five improvements, from one-evening client fixes to the prize chest that would make a stranger's basement worth raiding.
Twenty-five years of player-authored dungeon systems tell one story in three acts. Act one: systems that let authors mint server rewards melted down. City of Heroes' Mission Architect produced three-hour max-level farms within days of launch in 2009; EverQuest 2's Dungeon Maker had experience surgically removed from it in 2014; Neverwinter's Foundry lost its daily rewards in 2015 and was then killed entirely in 2019, when the studio admitted the knowledge required to maintain its bespoke editor pipeline no longer existed. Act two: systems that fixed farming by isolating rewards died of irrelevance instead. EQ2's play-as-an-avatar phase, where your real character never entered and a token store held the payout, proved that nobody grinds content that does not touch their real character. Act three: the survivors externalized cost and made the dungeon itself the economy. Neverwinter Nights persistent worlds are still setting concurrency records twenty-two years on because the community owns everything; and Old School RuneScape's player-owned-house dungeons have run un-nerfed for twenty years because the monsters are furniture bought with gold, the treasure chest holds the owner's stake, and the server mints nothing inside.
SotA's system already made the two hardest calls correctly. Dungeons are built from live-game primitives: the pieces are decoration items, the layout lives in the same storage as house deco, the rooms are real scenes. There is no bespoke editor to rot, which is the exact failure that killed the Foundry. And totem spawns were pre-emptively tuned to modest loot, so the farm meltdown never happened here. But the system stopped at the cost structure and never shipped the social loop. There is no way to find a dungeon, no reason to enter one, no record that anyone ever did, no map of your own halls, no traps, six creature families, and an acquisition grind that punishes the few who build. Measured against the genre, our gaps cluster in three places:
Twenty-five proposals follow: six client-only quick wins, eight core upgrades, and eleven server-backed flagships. The flagships share one design doctrine, taken from the only system that has survived two decades untouched: status pays the author, players pay each other, and the server mints nothing. Every server-backed card rides the two dungeon operation groups that already exist; none of them adds new network plumbing categories.
| # | Proposal | Tier | Size | Server | One-line pitch |
|---|---|---|---|---|---|
| S1 | Mode switch from inside Built | 1 | S | reuses op | flip Adventure and Design without walking out |
| S2 | Entrance stats tooltip Built | 1 | S | one read op | the doorway tells you what is down there |
| S3 | Placement denial reasons Built | 1 | S | none | say exactly why a room will not attach |
| S4 | Points follow the lot Built | 1 | S | none | upgrading your lot upgrades your dungeon budget |
| S5 | Shareable generator seeds Built | 1 | S | none | type a friend's seed, get their layout |
| S6 | Sample dungeon showcase | 1 | S | none | curated teaching dungeons behind every entrance |
| S7 | Live dungeon map Built | 2 | M | none | the M key finally works in your own dungeon |
| S8 | Server-enforced budget Built | 2 | M | existing ops | the point limit becomes real |
| S9 | Encounter variety pass | 2 | M-L | data | new totem families and mixed spawn rooms |
| S10 | Essence and totem retune | 2 | S-M | data | totem costs a player can actually reach |
| S11 | Battle statues | 2 | M | data | cosmetic fighters in any decorated room |
| S12 | Decorating spawn rooms | 2 | M-L | none | fights and furniture in the same room |
| S13 | Blueprint acquisition pass | 2 | S-M | data | base blueprints on booksellers, exotics stay rare |
| S14 | Trap and trigger kit v1 Building | 2 | L | none | plates, darts, gas, sealing doors, wave arenas |
| S15 | Clear-check publishing Built | 3 | L | publish flag | open to the public only after you beat it yourself |
| S16 | Entrance plaque and record Built | 3 | L | dungeon record | first clear, record time, visits, on a plaque forever |
| S17 | Two-axis post-run rating | 3 | M-L | record fields | one like plus one word: Fun, Deadly, Clever, Grand |
| S18 | The Delve Board | 3 | L | listing fields | an in-game directory of open dungeons |
| S19 | The author's ledger | 3 | M-L | record fields | who came, how far they got, where they died |
| S20 | Owner-staked prize chest | 3 | XL | escrow | stake your own gold; first to clear it keeps it Built |
| S21 | Time-attack medals | 3 | M | via S16 | author time seeds bronze, silver, gold |
| S22 | Featured dungeon rotation | 3 | S-M | via S18 | a human picks the week's dungeon, not an algorithm |
| S23 | Dungeon build jams | 3 | S | process | monthly theme, winners become sample dungeons |
| S24 | Dungeon Master mode Building | 3 | L | existing ops | the owner runs the dungeon live for guests |
| S25 | Piece-maker credit | 3 | S-M | via S16 | crafters' names ride their pieces into famous dungeons |
The R61 player-dungeon system is one of the deepest pieces of housing tech in the game, and this document deliberately builds on it rather than re-proposing any of it. What exists today: a dungeon entrance placed as lot or basement decoration; six foyer types; roughly 130 room and passage pieces across five material sets, including secret breakaway walls and encounter rooms modeled on famous dungeons of Novia; a doorway-click builder with a live snapped preview that refuses to orphan rooms from the foyer; a dungeon point budget by lot size, upgradeable with gold or Crowns; encounter totems in six creature families with tiers up to boss level; Adventure Mode with an occupant-safe countdown and reload; a Random Dungeon Generator chest that assembles a seeded layout from whatever pieces you own; an open-PvP flag with an honest warning on the doorway tooltip; custom dungeon names that always display beside the owner's name; and per-room baked pathing stitched together at runtime so the monsters can actually chase you through whatever you built.
A dungeon is decoration all the way down. The entrance is an interactive deco object on the lot; entering pushes the foyer scene with the owning lot carried along; the placed pieces load as deco items whose connection sockets hash into a room graph; the builder window validates a snapped preview against the piece sets, the budget, and a walk-back-to-the-foyer check before committing; and flipping Adventure Mode broadcasts to everyone inside, rebakes the stitched navmesh, and lights the totems, which pick the spawn table for their creature family and tier. Every proposal below extends one of these nodes. None forks the pipeline, and that is the system's crown jewel: there is no bespoke editor to maintain, which is the exact cost that killed the genre's biggest attempt.
flowchart LR A["Dungeon entrance\nlot decoration"] --> B["Entrance window\nenter / options"] B -->|Enter| C["Foyer scene loads\nwith the owning lot"] C --> D["Pieces load as deco\nconnection sockets"] D --> E["Room graph\nsnap + reachability"] B -->|owner doorway click| F["Piece picker\nvalidate + commit"] B -->|owner toggle| G["Adventure Mode\nbroadcast"] G --> H["Runtime navmesh\nstitch"] G --> I["Totems light\nspawn family + tier"] style F stroke:#c9a55a,stroke-width:2px style G stroke:#c9a55a,stroke-width:2px
Player housing invented the itch; player dungeons are the genre repeatedly trying to scratch it. Ultima Online proved players will pour years into spaces they own and show off, and its design contests, where winning house designs became official content, remain the cheapest author payment ever devised. Neverwinter Nights handed the whole cost structure to the community and became immortal. City of Heroes put the authoring tool inside a live MMO and learned, within days, what happens when authors can tune the rewards. EverQuest 2 built the closest system to ours, collect creatures and place them in rooms, and killed it twice, first with reward isolation and then with reward removal. The Foundry scaled quest authoring further than anyone and died of its own bespoke editor. Landmark shipped magnificent build tools with no audience loop and lasted seven months. And quietly, for twenty years, Old School RuneScape's construction skill has run dungeon rooms, monsters as gold-priced furniture, and an owner-staked treasure chest without a single economy scandal.
Old School RuneScape: the whole reward doctrine. Guards are bought with gold on a curve from cheap to legendary, the treasure room chest holds the owner's own stake, challenge mode arms it, and house parties made the space a social institution. Twenty years, zero nerfs, because the server never mints anything inside. Super Mario Maker 2: the clear-check (you must beat your own course to publish it), First Clear and World Record names on every course page, fixed tags, and difficulty from real clear rates rather than author claims. Trackmania: the author's validation run seeds the bronze, silver and gold medal times automatically. Meet Your Maker: the raid ledger, every visitor's death and path reported back to the builder, was the strongest builder-retention ritual the genre has produced; its economy, paying builders per raider death, bred degenerate kill-boxes and killed the game inside a year. Neverwinter Nights: the DM client, a human running the dungeon live, is why its persistent worlds outlived four generations of MMOs. Dreams and LittleBigPlanet: editorial picks surface quality; algorithmic popularity gets farmed by reciprocal rings at any population size. Solasta: place-and-configure activators deliver most of trap scripting's value with none of free scripting's moderation and maintenance bill. City of Heroes and EverQuest 2: the two cautionary tales, author-minted rewards and isolated rewards, which between them bound the safe design space every card in this document lives inside.
| Capability | NWN | CoH MA | EQ2 DM | Foundry | OSRS POH | SMM2 | MYM | SotA today | SotA proposed |
|---|---|---|---|---|---|---|---|---|---|
| In-game authoring (no separate tool) | ○ | ● | ● | ○ | ● | ● | ● | ● | ● |
| Dev-balanced pieces (no stat authoring) | ○ | ○ | ● | ◔ | ● | ● | ● | ● | ● |
| Trigger and trap wiring | ● | ◔ | ◔ | ◔ | ◔ | ● | ● | ○ | ◔ S14 |
| Encounter placement | ● | ● | ● | ● | ● | ● | ● | ◔ | ● S9 |
| Publish gate / clear-check | ○ | ○ | ○ | ◔ | ○ | ● | ◔ | ○ | ● S15 |
| In-game discovery / directory | ◔ | ● | ● | ● | ◔ | ● | ● | ○ | ● S18 |
| Rating and feedback | ○ | ● | ● | ● | ○ | ● | ● | ○ | ● S17 |
| Author telemetry (ledger, death map) | ◔ | ◔ | ◔ | ◔ | ○ | ◔ | ● | ○ | ● S19 |
| Runner reward without economy risk | ◔ | ○ | ○ | ○ | ● | ● | ◔ | ◔ | ● S20 |
| Creator reward as status | ● | ● | ◔ | ● | ● | ● | ● | ○ | ● S16 |
| Timers, records, medals | ○ | ○ | ○ | ◔ | ○ | ● | ◔ | ○ | ● S21 |
| Live DM / event layer | ● | ○ | ○ | ○ | ◔ | ○ | ○ | ◔ | ● S24 |
| Low moderation burden | ● | ○ | ◔ | ○ | ● | ◔ | ◔ | ● | ● |
| Built from live-game primitives | ◔ | ◔ | ● | ○ | ● | ● | ● | ● | ● |
● native · ◔ partial · ○ absent. Reading the matrix: SotA today is already strong on the two columns that killed everyone else, moderation burden and the primitives-based stack, and weak only on the audience loop. This catalog is an audience-loop program, not a rebuild.
Built 5 August 2026. Standing in a dungeon you own, a small panel in the top right corner of the screen names the current mode and offers the switch to the other one. It confirms first, warns that the dungeon will reload, and then runs the same switch the entrance window offers, countdown and all. Visitors never see it. Not on the test shard yet.
Owners must currently walk out of the dungeon to flip between Adventure and Decoration Mode, so every spawn-tuning iteration is an exit and re-enter cycle. The server operation already exists, already checks owner permission, and the countdown-and-reload path already handles people standing inside when the mode flips. This card only adds a place to press the button from down below: an owner-only toggle on the foyer, or the existing options menu surfaced from inside.
Seen in: every level editor with a play-test button; Mario Maker's instant edit-and-play flip.
The client call path (PersistenceService.SetDungeonAdventureModeOption) and the server validation and broadcast already exist; the occupant-safe reload coroutine on the plot handles mid-fight visitors with a countdown message. Work is UI-only: an owner-gated foyer interactive or a context action registered while inside the dungeon scene, calling the same op.
Built 13 August 2026. Hovering a dungeon doorway now reads out what is behind it: how many rooms, how many encounters, which materials it is built from, and whether Adventure Mode is switched on. The same rows appear in the window the doorway opens. The counts describe the dungeon as it stands, so a room added this afternoon is on the door this afternoon. One correction to the card as written: this could not be client-only. A visitor standing on the lot has none of the dungeon loaded, so the counts are worked out on the server from the placed pieces and ride the record the doorway already asks for. The same pass fixed the entrance plaque, whose first-clear and record lines had been withheld at every doorway since they shipped, for the same reason. Not on the test shard yet.
The doorway tooltip already shows the dungeon's name, its owner, and a PvP tag. Extend it with what the client already knows: piece count, encounter rooms, whether Adventure Mode is live, and the material themes inside. A visitor should be able to read the door before opening it.
Seen in: Meet Your Maker outpost cards; Mario Maker course info panels.
Built as a shared layout summary that walks the dungeon magazine's item stores where the record handler already loads the plot, counting rooms, encounters and material families. The three counts travel on the record reply as new optional fields, so an older client and an older server both keep working. Nothing is stored: a banked count would only be as fresh as the last clear, and the whole point is the dungeon as it stands. The inventory walk matches the one the S8 point budget enforces, entry for entry, so the doorway cannot advertise rooms the budget does not believe in. Encounters are placed totems plus the room variants that ship with their fight; materials come from the same name parser the piece picker uses, now shared rather than duplicated. The plaque's size check moved off the viewer's own count, which is always zero at a doorway, and onto the server's, which is the harder number to fake.
Built 5 August 2026. When a piece will not attach, the picker now says why in one line under the point rows: the piece belongs to a set the doorway does not accept (naming both sets), it would overlap another room, it would cut a room off from the foyer (naming the room), or it would go over the point budget (with the numbers). The Accept button still disables; the line explains it. Not on the test shard yet.
When a piece will not place today, the preview just stays red. The validator already knows exactly why: the piece belongs to a different material set than the doorway accepts, it would overlap an existing room, it would cut rooms off from the foyer, or it would break the point budget. Say so, in one line, in the picker. Builders should never have to guess which rule they hit.
Seen in: the decorating page's denial-reasons pattern, applied to the room graph.
The picker's validation pass computes four distinct failure causes but collapses them into one boolean. Thread the cause out to a message line; strings localized up front. The reachability wording reuses the walk-to-foyer check's result ("Placing this would cut the Crypt off from the foyer").
Built 5 August 2026. A dungeon's point budget now reads the lot it stands on instead of the lot it was founded on. Upgrade from a Row lot to a Castle and the budget rises to match. Downsizing keeps every room already placed and only pauses new additions, and the piece picker names how far over the limit the dungeon is until it is back under. Not on the test shard yet.
Your dungeon point budget derives from your lot size, but moving the dungeon to a bigger lot does not recalculate it: the budget stays frozen at the size of the lot where the dungeon was founded. Fix the derivation to read the current lot. Move to a Keep lot, get a Keep budget. Shrinking moves keep every placed piece and simply block additions until you are back under, the same over-limit convention the deco system already uses.
Seen in: RuneScape houses, where an upgrade unlocks rooms immediately.
The limit is entirely client-computed and two call sites disagree on the source (one reads the persisted magazine shape, one the parent plot type), which is the bug's shape. Derive both from the current parent lot shape. S8 later mirrors the same rule server-side; landing S4 first keeps the two implementations identical.
Built 5 August 2026. The chest's build confirmation now has a seed box. Leave it blank and the roll is random, the same as before. Type a number a friend gave you and the build uses it, so matching piece collections raise the same halls. Letters and other non-numbers switch the Build button off until the box is fixed or cleared. After every build the chat summary prints the seed the roll used, in plain digits you can retype. Not on the test shard yet.
The Random Dungeon Generator already runs from a seed. Print it when the build finishes, and accept a typed seed in the chest's confirm dialog. Two players with similar piece collections can trade layouts; a lucky roll becomes a thing you can give away. Sample completion line: "Build complete. Seed 88412. Share it so a friend can raise the same halls."
Seen in: roguelike daily seeds; Minecraft world seeds as social objects.
The layout core is already seeded, deterministic, and covered by EditMode tests; the chest interaction gains a seed field and the completion toast gains the number. Different piece inventories produce different results from the same seed; the message sets that expectation. Totem auto-placement and how the chest itself is acquired stay with the generator's own roadmap follow-ups.
Every entrance already has an Enter Sample Dungeon button and a per-entrance sample override, and five dev-built demos ship today. Add three or four new curated samples that each teach one technique: secret doors done well, encounter pacing, a fifty-point starter buildable entirely from bookseller recipes. Then make the sample rotation the landing pad for build-jam winners (S23), the Ultima Online contest pipeline where player work becomes official content.
Seen in: UO house design contests; Mario Maker's first-party courses seeding the course browser.
Content-only: new sample scenes registered to the existing per-entrance override. Each sample gets one plaque-style sign inside naming the technique it demonstrates.
Built 5 August 2026. Pressing M inside a player dungeon now shows a 3D model of the dungeon assembled on the spot from the rooms placed there. Each room draws as a translucent block, a gold beacon marks where you stand, the floor you are on is highlighted, and rooms that cannot reach the foyer draw in grey so a disconnected wing stands out. The map updates as rooms are added or removed and keeps your camera where you left it. No fog of war in this first version, as planned. Not on the test shard yet.
Player dungeons are the only combat spaces in the game with no M-key map, and owners of large builds famously navigate their own halls with camera tricks and a fan-made web mapping tool. The fix is structural: the client already knows every placed piece's bounds and how the rooms connect, so the map can be assembled live from the room graph, the same way the monsters' pathing already is. Version one ships without fog of war and reuses the 3D map rig the static dungeons got this summer.
Seen in: every dungeon crawler since Dungeon Master; SotA's own 3D maps for 58 static dungeons.
Build the map model from placed piece bounds plus the connection graph's position-hash adjacency, rendered through the existing 3D map rig (DungeonMap3DView) with a new live mode in the map-mode model. This is exactly the seam the scene-maps epic reserved for player dungeons; coordinated with that epic, not a fork of it. Fog is deferred because fog storage keys on scene name and every dungeon on the same foyer type shares one, so per-dungeon fog needs a key extension. Proxy-box rooms read blockier than baked models; acceptable for v1 and consistent with the fan tool players already praise.
Built 5 August 2026. The server now checks the dungeon point budget on every piece placed, using the same math the client shows in the picker. The base numbers per lot size moved into server configuration, so the client reads them from the server and future tuning needs no client patch. Dungeons that were already over the old limit keep everything they have; they just cannot add more until they are back under. Replacing a piece at the limit still works: the swap now frees the old piece's points before the new piece lands.
The dungeon point limit is enforced only by the client today; the server stores whatever the placement operations deliver. Mirror the budget math server-side and validate on every piece add, and move the base numbers into server configuration so tuning stops requiring a client patch. Nothing changes for honest players; the ceiling simply becomes real, which matters before anything competitive (records, prizes) stands on top of it.
Seen in: standard server-authority hygiene; Mario Maker validates uploads server-side.
Port the pure limit math into the shared code both sides compile (the shared-source mirror makes this near-mechanical) and validate in the plot-content piece-add path; configuration joins the existing dungeon rules config section. Legacy over-limit dungeons (from the frozen-budget era) grandfather as read-only-over-budget: additions blocked, nothing ever truncated.
Six creature families, one family per spawn room, is the builders' third-loudest complaint. The encounter chain is data all the way down: totem types, recipes, and per-tier spawn tables. Add three or four new families with existing world art and essence sources, skeletons, zombies, wolves, bandits, and widen the generic encounter rooms to accept any family so a builder can theme a crypt as a crypt.
Seen in: EQ2 Dungeon Maker's collect-the-creature loop, its one proven positive; RuneScape's ladder of purchasable guards.
New totem family enums, recipes cloned from the 36 existing totem recipes, and level-banded spawn assets per family, all respecting the modest-loot doctrine. The room's slot filter already supports multi-family acceptance (or-slots exist today); widening the generic rooms is data. The existing spawn-table ratchet tests are the regression net.
A top-tier totem costs 25,600 essences to combine, and converting a single spawn room to a decorated room costs 500 Crowns. The community's verdict on the grind is not printable in a family game. Flatten the combine curve so the top tier is a few hundred hours of themed hunting rather than thousands, make totem salvage return essences, and give the Crown-only room conversions a gold-and-materials recipe path. The totem ladder is this system's long-term sink and should stay expensive. It should stop being unreachable.
Seen in: RuneScape's guard ladder, a curve from cheap to legendary rather than a wall.
Recipe quantity edits on the existing totem recipe set plus essence drop-rate tuning. Existing top-tier owners get a one-time essence rebate on the delta so nobody's investment is devalued silently.
The Lich Fighter Spawner Statue, a decoration that spawns a hostile, no-loot, no-experience mimic in any decorated room, is the community's favorite trick for making dungeons feel alive, and event hosts already stack them. Formalize the trick: one spawner statue per creature family, placeable in decorated rooms under the normal deco budget, spawning cosmetic fighters that drop nothing and grant nothing. Combat as decoration, immune to the farm problem by construction.
Seen in: RuneScape's inert guards that flip live in challenge mode; the community's own statue stacking.
The existing statue is the template: per-family variants bound to zero-loot, zero-XP spawn assets, gated to dungeon scenes the way dungeon-only deco spawners already are. The per-lot NPC cap already bounds the perf worst case.
Today a room either fights or it furnishes: spawn rooms may hold zero decorations, decorated rooms may hold zero spawns, and converting between them destroys one half. It is the single biggest creative-freedom complaint builders have. Let decoration into spawn rooms under a reduced cap, no containers, with keep-clear zones around the spawn points so creatures can never be walled in. A crypt can finally have coffins in it while the skeletons climb out of them.
Seen in: RuneScape rooms, where traps, guards and furniture coexist; Meet Your Maker, where traps and decor share every tile.
Relax the room-type exclusivity flag under a reduced deco allowance; keep-clear volumes validate at placement and the runtime navmesh scan already refuses unreachable spawns. The spawner-in-wall regression suite replays validation over every kit room. The decorating page's D19 owns generic deco behavior inside dungeons; this card changes only the room-type rule, and the two land coordinated.
Every dungeon blueprint is a rare drop from a handful of zones and bosses; a veteran can spend an evening failing to buy the most basic entrance recipe because no bookseller carries it. Put the base set, one full hall-and-room family plus the starter entrance and foyer, on booksellers at honest gold prices. Keep the famous-dungeon encounter rooms and themed exotics as the trophies they are, and add bad-luck protection so the drop-gated tier is a hunt, not a lottery. Also make consumed store-bundle blueprints re-grantable to the buying account, closing a long-standing support sore.
Seen in: RuneScape, where every base room is purchasable and the prestige items are earned.
Bookseller inventory data plus pity counters on the drop-gated tier; store-bundle grants keyed to the purchasing account. Stagger the bookseller additions so existing blueprint resale values deflate gently rather than overnight.
Part-built, 14 August 2026. Nine pieces exist and can be placed in a dungeon and connected to each other. Harmless: a Pressure Plate, a Shrieking Gargoyle that shrieks and takes no health, a Snuffer Brazier that puts out every torch in its room, a Message Plaque that speaks a line its owner wrote, a Sealing Door that shuts on its trigger and opens again when it clears, and a Boss Door that only ever opens. Damaging: a Dart Wall, a Flame Jet and a Spike Floor, all at the gentlest tier. Right-click a placed piece and choose Connect To and it lists the other pieces standing in the same room; effect pieces get a second menu for how long they wait and whether they ready themselves again, and the dungeon builder panel reports the first real problem with a room's wiring and names the piece. A wave spawner is written as well, waking the creatures a room's own totems already seated rather than adding any.
The damaging pieces are bounded by a running account rather than a per-hit number alone: one trap over ten seconds, one room over ten seconds, and a whole dungeon over a minute all have ceilings, so a corridor of the gentlest traps cannot do what none of them could do alone. A player who simply stands still in one survives the minute. Still to come: a gas vent, which needs art. Not on the test shard yet, and nothing grants the pieces yet.
The kit has no interactive vocabulary: no pressure plates, no darts, no gates that close behind you, no wave arenas. Every mature dungeon-building game converges on the same ten families of trigger and effect, and none of them requires scripting. Ship configured activators: trigger pieces (pressure plate, tripwire, lever) wired by pointing them at effect pieces (dart wall, gas vent, sealing door, message plaque, wave spawner using your existing totem creatures, a boss door that opens when the room is cleared). Damage is capped per effect so traps wound and frighten rather than one-shot, and plaques pass the same name sanitation every player label does.
Seen in: Solasta's place-and-configure activators; Minecraft's command-block grammar; Meet Your Maker's trap set.
Trigger and effect are deco-item components using the same parent-piece association the totem already uses; wiring is point-at-target in decoration mode, the same interaction grammar as the doorway picker. Effects run under adventure-mode gating like spawns do; state lives in the room's existing master-client model. The v1 line is fixed pieces with enum parameters, nothing free-form: the genre's scripting systems all ended in moderation or maintenance debt. Lethal traps should not outrun S15's beatability gate; land them in the same release.
Built 5 August 2026. The publish switch lives in the entrance window's options menu. Publishing is refused until the owner has cleared the dungeon's current layout in Adventure Mode; a chat line explains what is missing. Editing the dungeon's contents or swapping the level closes it to the public again until the owner re-clears. One correction from the card: a clear means defeating every monster the dungeon spawns in one Adventure Mode visit, and the announcement reaches everyone inside the dungeon at that moment. Not on the test shard yet.
A dungeon becomes Open to the Public only after its owner has beaten it, foyer to final encounter, in Adventure Mode, with the run time recorded. One mechanic buys three things: every public dungeon is provably beatable, troll design is capped at what its own author can survive, and the author's time becomes the seed for the medal ladder. Editing the dungeon retires the publication until it is cleared again. Private dungeons behave exactly as today; nothing is taken from anyone. Sample string: "This dungeon is not yet open to the public. Clear it once in Adventure Mode to publish."
Seen in: Mario Maker's Clear Check; Trackmania's validation run; Mighty Quest's self-validation.
A publish flag and author-time fields beside the existing per-plot dungeon options; one new operation riding the existing plot-content group. Clear detection consumes the encounter-clear events rooms already emit. A server-side plausibility floor (minimum elapsed time against piece count) filters lazy claims; everything a clear unlocks stays modest by design. Recommended after S8 so publication implies a budget-honest dungeon.
Built 5 August 2026. Every player dungeon now keeps its record: total visits, distinct adventurers, clears, the permanent First Clear, the record time and its holder, and the author's own best. The entrance hover tooltip and the entrance window both show the plaque. First Clears and record breaks are announced to everyone inside the dungeon. Record lines appear on published dungeons with at least 3 pieces; visits count everywhere. One correction from the card: the announcement is heard inside the dungeon in this first version, and guild chat comes later with the medals work. Not on the test shard yet.
One server record per published dungeon: total visits, distinct visitors, clears, the permanent First Clear name, the record time and its holder, the author's own time. Surface it as a plaque on the entrance and in the entrance window. In a small world the currency is names, and the plaque is what makes building worth it: your dungeon becomes street-visible status furniture with other people's names engraved on your doorstep. Record breaks announce in local and guild chat.
Seen in: Mario Maker's First Clear and World Record names on every course page.
A new dungeon-record store keyed by the scene-and-plot identity the game server already carries for dungeon rooms; visit increments hook beside the existing recent-scene tracking; one read operation rides the existing dungeon op group. Records display only above a minimum piece count so one-corridor dungeons cannot farm the ladder. One migration-ledger row for the new collection and its indexes.
On leaving a published dungeon after a qualifying run, one small prompt: a Like, plus one word for the builder, Fun, Deadly, Clever, or Grand. Danger is never self-declared; it is computed from real death and clear data. Two axes let a brutally hard but fair dungeon build a reputation as Deadly instead of eating angry thumbs-downs, and the descriptor tallies feed monthly titles. One rating per account per dungeon, and only after a real run.
Seen in: Meet Your Maker's accolades plus computed danger rating; Mario Maker's like and difficulty votes.
Rating fields on the S16 record; one write op on the existing dungeon group; the prompt hooks the scene-exit path when popping a dungeon scene after a qualifying run. Ratings weight toward accounts with play history, the standard defense against reciprocal-voting rings, and the featured rotation (S22) never keys on raw score.
An in-game directory of published dungeons, opened from a placeable tavern board, the town crier, and a map button: Newest, Top Rated, Never First-Cleared, and Curator's Picks, filtered by a small fixed tag vocabulary the author chooses at publish (Combat, Puzzle, Maze, Story, Showcase, Boss, Trap Gauntlet, Group). Selecting an entry marks the entrance on your map. A Surprise Me button picks a random dungeon that has never been first-cleared. Discovery is this system's hardest problem, and at our population a good list beats any algorithm.
Seen in: Portal 2's queue, the lowest-friction discovery pipe ever shipped; Mario Maker's fixed tag rows.
Reads the S16 records plus publish flags through one list op on the existing dungeon group; wayfinding reuses the lots-on-map marker path that shipped this month. Clear rates display next to tags so mis-tagging is self-evident. Launch seeded with the S6 samples and a few dev-built entries so the board is never empty. The board is deco, so towns can adopt it; any web rendering of the same directory belongs to the companion-site program, not this page.
The dungeon reports back. An owner-only log of recent runs: who came (or "A wandering adventurer" for players who prefer privacy), how far they got, where they died, how long they took, what rating they left. Deaths draw as markers on the S7 live map, your own private heat map. Builders in every game that shipped this describe checking it as the daily ritual that keeps them building.
Seen in: Meet Your Maker's raid replays and death maps, the genre's strongest builder-retention hook.
A capped ring buffer of recent runs on the S16 record, fed by death and progress events the room already produces; one owner-gated read op on the existing dungeon group. No replay capture: markers and times deliver most of the value at a fraction of the cost. Anonymous by default for non-friends, opt-in naming.
The flagship, imported whole from the only player-dungeon economy that has survived twenty years untouched. A Prize Chest piece the owner stocks with their own gold or items. It arms only in Adventure Mode, only in a published dungeon, and the first visitor or party to reach it after a clean run claims the stake. Restocking is throttled. Rewards flow from player to player; the server mints nothing, so there is nothing to exploit, and the chest is a voluntary gold sink at whatever scale the owner's pride sets. This is the card that answers the community's oldest question about player dungeons: why would anyone go in? Sample broadcast: "Elrik has staked 50,000 gold in the Sunken Vault. Unclaimed for 6 days."
Seen in: RuneScape's treasure room and challenge mode, un-nerfed since 2006. Every server-minted alternative in the genre melted down.
Built 2026-08-13. The Dungeon Prize Chest is in. You place one inside your own dungeon and put your own gold in it. While the dungeon is open to the public and in Adventure Mode the stake is committed, so you cannot pull it back out, and the first visitor who clears the dungeon takes the lot. Close the dungeon and you can take your stake back whenever you like. The entrance advertises the prize, so people can see there is something worth the trip before they walk in, and the dungeon hears about it when someone wins. Nothing is created out of thin air: the gold leaves one purse and lands in another. Two things came out differently from the card. One winner takes the whole stake rather than a party splitting it, and the stake is gold rather than gold or items, because a player's gold is a balance rather than something you can drop in a box. There is a cap on how much can sit in a chest, and a cooldown before a claimed chest can be stocked again.
Stake and claim operations ride the existing plot-content group with the stake held in server escrow; claim validation is entirely server-side (published, adventure mode, armed, clean run) and reachability is guaranteed by S15's clear-check. The claim path must be atomic, one claimant, escrowed funds, no duplication on disconnect, which is the real server engineering and the reason this card is XL. Claiming your own stake through an alt is legal by design: the owner loses only their own gold. Economy events log through the same channel the existing dungeon purchases do. One detailed migration-ledger row.
The author's clear-check time automatically seeds Bronze, Silver and Gold thresholds. Your personal best lives on the plaque; a top-ten board lives on the record; new records announce in local and guild chat. Friend-scoped comparison beats a global ladder at our population: global boards go stale, friend boards stay contested forever.
Seen in: Trackmania's author medals; Mario Maker's world records.
Pure derived data on the S16 record; no new operations. The plaque shows the author's level and party size beside the author time, the standard context against sandbagged slow times. No ghost replays in v1; S19's death markers already deliver the learning value.
Once a week, a human picks one to three featured dungeons: a ribbon on the plaque, the top of the Delve Board, a mention in the weekly notes. Editorial curation is the entire answer to ranking abuse at our scale; the genre proved that votes surface farms and front-page algorithms get gamed at any population, while a named human's pick stays trustworthy. Simple published criteria; never the same author two weeks running.
Seen in: Dreams' Mm Picks; City of Heroes' Developer's Choice; the Foundry's spotlight.
A featured flag on the S18 listing; the selection is a weekly operations task, not code.
A monthly themed contest: Undead Crypt, No-Combat Puzzle, Fifty Points from Bookseller Pieces Only. Deadline, judging, winners featured for a month and immortalized as S6 sample dungeons, every entrant gets a trophy piece. The bounded prompt is also the best onboarding the kit could have: a theme teaches the pieces better than any manual. This is the Ultima Online contest pipeline, the strongest author payment the genre ever found: your work becomes part of the game.
Seen in: UO's house design contests; Dreams' community jams.
Process plus the existing per-entrance sample override plus one trophy deco archetype. Scale the ceremony so five entries still feels like an event; announcements ride the existing message-of-the-day and blog channels.
Part-built, 14 August 2026. A dungeon's owner or co-owner can start a live event from the same row they switch the dungeon's mode on. While one runs, nothing is recorded, and that is the server refusing rather than the client being polite, so no visit, clear, first clear or record can come out of a run a person was animating. The console lists the wired channels in the room the host is standing in and fires them by hand, speaks a line from a curated set through the nearest creature, and opens a reward chest the owner locked and filled beforehand. An event needs the dungeon to be in Adventure Mode already, and it ends when the person running it leaves. Everyone inside is told when one starts, and somebody who walks in during one is told on arrival. Still to come: a warning before you walk in, which needs the mode to be visible from outside the dungeon. Not on the test shard yet.
An owner-toggled live-event mode for a published dungeon: fire your S14 wave and trap triggers by hand, speak a boss line from a curated set, drop a staged reward you provided, while medals, records and the prize chest are switched hard off so the mode can neither farm the ladder nor grief it. This is Neverwinter Nights' DM client shrunk to a housing MMO, and it is the highest-leverage card per developer-hour in this document: a human animating the space turns a one-visit museum into a repeatable guild night.
Seen in: NWN's DM client, still running twenty-year-old persistent worlds; Divinity's Game Master mode; RuneScape house parties.
A transient mode flag using the same broadcast pattern Adventure Mode uses, riding the existing plot-content group with no persistence. Manual firing is owner-only invocation of S14 activators; barks come from a fixed localized line library, no free text. The mode is loudly signposted at the entrance and exit is always available.
Dungeon pieces are crafted, and crafted things carry a maker's mark. Carry the mark through to the assembled dungeon: the plaque lists "Pieces by: Aria (12), Bren (4)", and a monthly most-placed-crafter title rides the rating data. Crafters gain a stake in other people's famous dungeons, and piece-crafting becomes a reputation market, which is the truest kind of economy this game has. Recognition only, never gold per placement; paying per event is the one lesson the genre wrote in blood.
Seen in: Dreams' automatic ancestry credits; SotA's own maker's-mark tradition.
Aggregated client-side from the maker's marks already on the magazine's items, surfaced through the S16 plaque payload; store-bought pieces credit "the Crown workshops". No schema change unless cached into the record, in which case it rides the S16 row.
The danger with twenty-five proposals is a game with twenty-five new widgets. The system has exactly two windows today, the entrance and the piece picker, plus the M-key map. The composition adds one new window and grows what exists: the entrance window gains a plaque header and an owner tab strip, the picker gains filter chips and one wiring row, the Delve Board arrives as the single new surface, and everything in-dungeon is a thin HUD layer. The plaque itself is one component rendered in three places: entrance window, Delve Board detail pane, and an optional physical plaque decoration by the door.
| Ledger, last 7 days | Result | Rating | Note |
|---|---|---|---|
| Aug 3 · Tamsin | Cleared 12:41 | Liked, Clever | New record |
| Aug 3 · A wanderer | Fell at Lich Vault, 9:02 | ||
| Aug 2 · Bowen | Cleared 14:58 | Liked, Deadly |
| The Sunken Crypt | Elyndra | Deadly | 92% |
| The Gilded Warren | Marcus | Fair | 88% |
| Halls of Rust | Petra | Brutal | 74% |
| Root Cellar Runs | Ansel | Easy | 95% |
| Phase | Ships | Server | Depends on | Why this order |
|---|---|---|---|---|
| 1 | S1–S6 in one client pass; S10 + S13 as the tuning companion | tuning data only | — | No dependencies, no server risk; every card retires a documented builder complaint. The existing audience stops bleeding. |
| 2 | S7 live map, S8 server budget, S9 + S11 encounter breadth, S12 spawn-room deco | S8 validation; spawn data | S7 before S19's markers; S8 before anything competitive | Dungeons become worth walking through, and the budget becomes trustworthy before records stand on it. |
| 3 | S15 publish, then S16 record, S17 ratings, S21 medals, S25 credit | one coherent record migration | S8 recommended first | The authorship spine: one schema wave, one ledger entry set, all ops on existing groups. Building gains a name and a scoreboard. |
| 4 | S18 Delve Board, S19 ledger, S22 featured, S23 first jam | listing fields | S15/S16; S19 wants S7 | The audience loop: discovery, the daily builder ritual, curation, and the contest pipeline. The ghost town gets traffic. |
| 5 | S14 trap kit, S20 prize chest, S24 DM mode | S20 escrow | S14 with S15's gate; S20 needs S8 + S15; S24 needs S14 + S15 | The flagships: dangerous halls, staked treasure, and a human running the night. The loop the genre's survivor proved out, complete. |
flowchart TD P1["Phase 1\nquick wins + tuning"] --> P2["Phase 2\nmap, budget, encounters"] P2 --> P3["Phase 3\npublish + record + medals"] P3 --> P4["Phase 4\nDelve Board + ledger + jams"] P3 --> P5["Phase 5\ntraps + prize chest + DM mode"] P2 -->|"budget is the trust floor"| P5 style P3 stroke:#c9a55a,stroke-width:2px style P5 stroke:#c9a55a,stroke-width:2px
| Card | New data | Network |
|---|---|---|
| S15 publish | publish flag + author time beside the existing per-dungeon options | one op on the existing plot-content group |
| S16 record | one dungeon-record entry per published dungeon (visits, clears, first clear, record) | one read op on the existing dungeon group |
| S17 ratings | rating tallies on the record | one write op, same group |
| S18 board | tags + listing fields on the record | one list op, same group |
| S19 ledger | capped ring buffer of recent runs on the record | one owner-gated read, same group |
| S20 prize chest | chest state + escrowed stake | stake/claim ops on the existing plot-content group |
| S21 medals | derived thresholds + top times on the record | none new |
| S24 DM mode | none (transient room state) | reuses the mode-broadcast pattern |
Every operation above joins one of the two dungeon-related operation groups that already exist and are already dispatched on the servers. No proposal adds a new network plumbing category, which removes the riskiest class of server change from the entire program. Schema changes land as one coherent migration wave in Phase 3 plus the S20 escrow in Phase 5, each with its migration-ledger entry in the same change.
The pure-logic pieces (placement denial causes, limit derivation, seed determinism, the map graph, medal thresholds, rating tallies, publish state machine, claim state machine, credit aggregation) are all EditMode-testable without scenes, extending the existing dungeon suites: the generator layout tests, the spawner placement ratchet, the totem spawn-table ratchet, and the entrance and map model tests.
Neverwinter Nights persistent-world histories and DM client documentation; City of Heroes Mission Architect launch coverage and the farm-nerf timeline; EverQuest 2 Dungeon Maker release notes and the 2014 producer's letter removing experience; Star Trek Online and Neverwinter Foundry retrospectives and the 2019 shutdown notice; Old School RuneScape wiki (Construction, dungeon rooms, treasure room, house parties); Super Mario Maker 2 course-world documentation (clear check, first clears, world records, tags); Trackmania medal system documentation; Meet Your Maker economy retrospectives; Dreams and LittleBigPlanet curation histories; Solasta Dungeon Maker documentation; community threads on visiting player dungeons, blueprint acquisition, spawn-room decoration, and the fan-made dungeon minimap tool. Full source list in the research notes accompanying this proposal.
Prepared 2026-08-04 against the unity-6.5 migration branch. Companion markdown: docs/plans/2026-08-04-player-dungeons-improvements-proposal.md. Twenty-five proposals: six client quick wins, eight core upgrades, eleven server-backed flagships, each grounded in the machinery that would carry it.
No account needed. Every proposal above also has its own Discussion box for talking about that one alone; this thread is for the plan as a whole. Comments are plain text, held to basic decency, and may be trimmed.