Hold The Horde — Regional Discovery Map

This page is a four-region, twelve-objective SMALL planning proposal. Region names and every objective placement after the selected opening are proposals, and found functioning parts are the acquisition default. The station shifts after every death; selected joins and completed work persist, while old routes become historical.

How to read dependencies in a shifting station

An objective dependency is a fact, carried object, completed change, or tested working rule required to make an action intelligible. It is never a promise that two rooms will be adjacent after death. When an eligible objective's room is not reachable in the current arrangement, the player pursues another eligible lead or returns. The intended ordering does not require repeated deliberate deaths solely to roll one exact target adjacency. Bearings and historical records guide investigation but reveal no unseen route.

Each objective below has one main dependency and a non-catastrophic alternative or honest deferral. Missing a clue, part, or survivor changes the evidence or cost, not whether the twelve-objective map can continue. Local joins bind only their selected room faces/groups. The entire-station connection is objective 12, deepest and latest, and is not a requirement for an apparent or deeper ending.

Region 1 — Dock Collar and Service Intake

This ship-rooted territory teaches carried versus installed property and gives later objectives portable prerequisites.

ID / layer	Player question and action	Entry condition; alternate or deferral	Useful return and later use	Pressure/build tradeoff; durable work versus knowledge
1. Collar isolation — opening	Can one Fork Relay (the splitter) power rack and shutter while the isolation bridge is removed, or should it remain carried? At this board, both outputs visibly accept the Fork's split. Complete fasteners and bring the bridge aboard.	Begins from the dock. The board selector can power the rack without Fork, so misunderstanding or leaving the part never blocks extraction.	Isolation bridge stops damaging back-feed and lets the ship report R-3's live bearing, never a route. Fork may be carried for builds or left operating the board.	Final work is noisy; Fork installation buys a blocked approach but costs time while the game runs and carried capacity later. Fasteners, bridge installation, board state, and Fork location persist; R-3 remains an observation.
2. Pace calibration — early	Does the rehab/service pace counter distinguish deliberate grounded travel from imposed motion of the same eligible subject? The astronaut walks across its marked interval, then stands feet-still on a service belt that carries the same astronaut across the same interval. Walking produces the pulse; belt motion does not.	Requires only reaching this placed calibration bay. If inaccessible, the player defers; a later pace counter can repeat the same comparison.	Found Stride Switch and a navigation coupler, the second ship assembly. Stride → Projector makes deliberate movement a build input rather than a damage tier.	Testing spends oxygen in a route crowd; carrying the coupler and part competes with a spare canister. The record supports this device's voluntary-versus-conveyor rule under the observed conditions, not every motion source or route's safety.
3. First local latch — early/middle	Which adjacent threshold faces accept the visible service pulse, and how can a mistaken join be released? Inspect, attempt one compatible join, then walk and test its network-side release.	Requires two currently adjacent, observed faces. This Collar instance has one wall-fixed diagnostic cradle whose charged output and compatible type are directly shown after objective 1. A mismatch refuses without consuming its current pulse; no compatible pair means honest deferral.	One persistent ship-rooted local connection and a recorded failed/successful signature comparison. Later it supports transformer reasoning.	The cradle supplies the local activation. A Fork can split only an existing input where both shown thresholds accept its shares; it is not itself a source. Power does not maintain the latch. The join persists; diagnostic power and the unbound neighbours beyond its group do not.

Region 2 — Clinical Exchange

This hospital territory ties one short human rescue to a first language test and a build-changing pressure objective. None requires the survivor.

ID / layer	Player question and action	Entry condition; alternate or deferral	Useful return and later use	Pressure/build tradeoff; durable work versus knowledge
4. R-3 crossing — early	Is someone alive, and is the present route short enough for one cooperative crossing? Personally walk airlock corridor → one junction → R-3 in one life; prepare air and hold that single junction.	Isolation bridge installed and the exact short current geometry personally walked. If absent, the scene waits; it never expands into an escort.	One person aboard only after that same person arrives and the airlock closes. Their historical testimony can help compare records, but no later objective requires it.	A canister and junction defence cost capacity/time. Completed preparation persists under rescue rules.
5. Builders' runs — early/middle, language 1	Do swapped first tokens track which of two visible mains ends in a room? Record the pump hall and cable junction, then physically compare them. Each room also contains one earlier A/B plate beside pulled or intact work and one builder workstation whose utterance address slot matches a token on that workstation's own fitting.	Requires records of both cases, not a fixed route between them. Either order works; one missing case defers the conclusion. A survivor may name crew roads but cannot translate.	Working prediction: the first token may name the main purpose. Later it helps choose a pressure control and identify related converter rooms. Its two A/B and two address observations are evidence for objective 9, not automatic conclusions.	The detour costs oxygen/exposure; journal comparison is paused but capture happens while the game runs. Records and notes persist. The result favors one reading without proving grammar, address meaning, or present connectivity.
6. Pressure ledger — middle	Can the language predictions identify sample versus purge and a likely redirected line, and is the shorter maintenance continuation walkable now? Test the controls and traverse or reject the prediction.	The instance must expose a genuinely reachable long approach in the current arrangement; if it does not, this objective defers. Visible gauges make language optional. Runs/pairs inform the riskier continuation. The long access is a placement obligation, not an every-shift guarantee.	Found Impact Dynamo plus pressure regulator core, the third ship assembly. Dynamo makes local Barrier/Projector impacts feed later storage; the ledger changes a later investigation or survivor account.	Holding a valve invites a local horde. Carrying both rewards may displace oxygen or salvage. Closed caps and freed objects persist; the observed destination/ledger are evidence, never proof of why transfers occurred.

Region 3 — Ascension Works

This territory layers older-looking mechanisms with the second alien language and a fourth ship function without confirming who built either.

ID / layer	Player question and action	Entry condition; alternate or deferral	Useful return and later use	Pressure/build tradeoff; durable work versus knowledge
7. Transfer pairs — middle, language 2	Do `OSH·a` and `OSH·b` mark source and receiving ends? Record three physically visible one-way transfers, predict one paired destination, and test a continuation. The three sites also supply three more pulled/intact A/B contrasts; one contains a third builder workstation/address-slot match.	Two pairs can support the directional hypothesis; the third is correction/redundancy and the route can be tested later. Without enough pairs, use the long route or defer.	A tested/corrected source–destination rule that can point toward objective 6 or the concrete boundary in 11. Arrival elsewhere is recorded as that outcome.	The uncertain continuation costs oxygen and may consume the trip's objective window. Records persist; material flow, repeated endings, and address matches never prove a walkable route, full translation, or speaker.
8. Balance and decontamination — middle	Can Phase-Balance plate order and a Breath Bank's travelling indicator predict a safe control window? Pin one rib or cap one vessel, then reuse the tell elsewhere.	No special part is mandatory: timing, hard cover, loose mass, and cutter checkpoints suffice. If the room is unreachable, another family instance can teach its narrower rule. An extra A/B plate here is redundant evidence, not required for 9.	Found Vector Nozzle and thermal exchanger core, the fourth ship assembly. Nozzle creates position/recoil choices; the completed cap/pin shortens a later traversal.	A large part plus assembly competes with oxygen/loot. Nozzle can move eligible mass but demands recoil clearance. Pins/caps persist; cadence knowledge travels, while the room's neighbours do not.
9. Regulator reply — deeper language preparation	Are A/B endings only a status display, or does seating the matching B-plate change the writing without touching the regulator? Pull R4, compare the new line, then test one valid and one invalid reply.	Objectives 5 and 7 already place five A/B contrasts and three builder workstation/address comparisons; objective 8's extra plate and the opening berth token are optional redundancy. The player must capture and compare them; the journal awards no learned flag. Missing the scar/plate leaves ordinary strip work and seals usable, and other occurrences can rebuild the correlation.	An interaction protocol: named work may advance a line, a matching B-plate may defer it, and invalid plates do nothing. Later it helps predict the inlet fitting at objective 11, never proving a mind or request.	Pulling is noisy and may yield a reusable Capacitor-family component under the found-parts default, but deferral can preserve an enabled mechanism. Completed pulls, seated plate, and written lines persist; interpretation remains a note.

Region 4 — Core Transfer Fringe

This is a late territory around—not an answer about—the central mystery. It contains an apparent ship-readiness opportunity, one deeper access problem, and the optional station-wide construction.

ID / layer	Player question and action	Entry condition; alternate or deferral	Useful return and later use	Pressure/build tradeoff; durable work versus knowledge
10. Ship readiness test — apparent layer	Do the four recovered functions let the ship present a credible departure preparation? Install bridge, coupler, regulator, and exchanger, then inspect a readiness report.	Objectives 1, 2, 6, and 8 supply the assemblies. Missing one leaves its named function incomplete, not a hidden failure. The player can continue exploring.	A visible preparation state and a clear list of missing functions, not automatic escape. If departure is later selected, this can become its preview boundary.	Four assemblies consume expedition capacity while recovered but become ship fixtures after installation. Installation persists; the report makes no duration, safety, or ending-truth claim.
11. Pressure-receiver iris — deeper layer	A sealed iris shows an empty receiving vessel, a low gauge, two builder-marked selectors and an inward flow arrow beside visitor `OSH·b`. Set one selector, observe flow, then hand-cycle the iris only after pressure equalises. Choosing the pump-purpose/inward selector fills the vessel and releases one passage to a deeper evidence cell.	Finding and physically reaching this assembly is mandatory; knowledge is not. Builders' run order predicts the pump selector. Visitor direction independently predicts feeding the marked receiver rather than purging it. Objective 9's current A-line naming the inlet is supporting evidence. Selecting purge/outward lowers the gauge, lights the vent and leaves the iris locked—evidence that corrects the choice without erasing work. Return to neutral and reverse it. A placed two-clamp mechanical equalisation line is the slower no-language alternative; its control/recovery access must be genuinely reachable in this boundary assembly, not guaranteed by an arbitrary shift.	One persistent opened passage and access to a deeper evidence site whose observations may revise the leading hospital/godhood hypothesis without settling it. No journal grades which clue was “enough”; each route guides the same physical manipulation and the gauge supplies the result.	Setting/cycling happens while the game runs and may be contested. A control build protects the panel; the two-clamp line costs more exposure/oxygen. Completed clamp and iris stages persist.
12. Connect every placed region — optional deepest/latest	Can every placed playable room/group become one ship-rooted traversable persistent network while every selected local join survives? Make/recover joins and verify physical inclusion.	Requires discovering all placed regions and the eventual join mechanism. Under the current mechanism, repeated refusals and converter observations from objective 3 support rare bridges; a different mechanism would replace that dependency. Never guaranteed by the next shift.	The entire-station construction itself, plus optional diagnostic evidence. It is not required for objectives 10 or 11 by default.	Each bind can cost travel, oxygen, and a current adjacency opportunity. Deliberate death may legitimately request another arrangement, but it is not a listed prerequisite. Joins persist; unobserved faces/routes remain unknown.

Two viable dependency orders

These arrows show prerequisites and chosen priorities, not physical paths. Current access can interleave eligible objectives; either order can defer R-3 until its short geometry exists.

Build-and-ship first: 1 → 2 → 3 → 4 → 8 → 6 → 10, then 5 → 7 → 9 → 11 → 12. The player first earns movement, impact, and positional parts and exposes the apparent-departure preparation, then turns to the two languages and deeper comparison. Objective 6 can be completed from gauges/long access before its language-supported shortcut is understood.
Language-and-network first: 1 → 5 → 7 → 3 → 6 → 9 → 4 → 8 → 11 → 2 → 10 → 12. The player takes riskier observations early, uses tested language predictions to shorten later work, and reaches deeper evidence before assembling all ship functions. The survivor arrives whenever the short crossing exists; their absence never blocks 5, 7, 9, or 11.

The only universal opening dependency is objective 1. Objective 10 needs the four named assemblies but is not a final ending. Objective 11 always asks for the same physical pressure-equalisation action; several evidence routes can make that action predictable, but none replaces manipulation. Objective 12 depends on complete placement and a selected joining rule, not on objective 10, rescue, treatment, or an ending. No required gate points back to its own reward.

One concrete later expedition — the pressure ledger

Back aboard, the player compares two builder-run records and several visitor transfer pairs. If the R-3 survivor is present, they may attribute the pressure-loop identifier to an old ward measurement; without them, the repeated identifier and live gauges still establish the objective, more slowly. Neither source reveals a route. The current airlock view offers one unwalked branch, so the player takes one fewer canister to reserve capacity for the suspected regulator core and an unknown part. That is a meaningful cost: less breathing margin now for the chance to bring both findings home.

At the loop, the builder rule predicts which marked control concerns pressure work, while the visitor source/receiving pair predicts that the line once ran toward a decontamination bay. The player records both claims before acting. They can take the current long approach, or test the maintenance continuation and accept that it may arrive elsewhere. At the valve, a local horde contests the hand-cycle. If objective 8 came first, a later-area 5 A / 5 C rig—Threat Sensor (1) → Pulse Projector (1) for ordinary automatic fire, plus Manual Trigger (1) → Vector Nozzle (2)—shoves eligible bodies into the announced draw lane; a strong execution can win cleanly. Without Nozzle, the four-space opening Sensor/Projector and Manual/Barrier rig can time the Breath Bank, protect one cap window, close that persistent cap, and withdraw. The journal pauses every threat and action while records are compared, and completes nothing.

On success, the player removes the Impact Dynamo from the sampling loop and frees the pressure regulator core. If both will not fit, they choose between immediate build change and ship readiness, drop another carried object deliberately, or return for the remaining item; death would bring everything still carried home but shift the station. Back aboard, the regulator advances objective 10's visible checklist, while Dynamo enables local impacts to feed storage rather than merely increasing damage. A present survivor may revise one attributed claim after seeing the ledger; without the survivor, the journal still records that the named transfer happened later than an earlier dated record. The expedition changes a build, a ship action, and one historical question without declaring why the transfer occurred.

Reuse and scope

The twelve objectives are cross-listed content. Objective 6 reuses the Language Expedition's runs/pairs, Traps' Breath Bank, a route crowd, the catalogue's Dynamo, one ship assembly, and a conditional survivor consequence. Objective 9 is the existing regulator-gallery chain rather than a new second-language system. Objectives 3 and 12 reuse the same face/join vocabulary at different depth.

Carried property returns on death; consumed or installed state does not duplicate. No person, treatment, whole-station connection, final departure, or story interpretation is mandatory here.

Open questions

These rules are untested — there is no evidence yet that the sequence is readable, varied, feasible, or supports any playtime ambition; numbers are proposals.

The actual shift/placement constraints. The shift rules are unresolved; whether reachability and pacing work under them is unproven, and unresolved placement may still make a new arrangement necessary for objective 12.
Whether the entire-station connection (objective 12) should relate to an apparent or deeper ending, and what it does to future shifts or endings; the final connection mechanism and full-station consequence.
What happens on death during the R-3 transfer. It is unresolved and cannot grant delivery, safety, or loss.
Selected cast and survivor outcome rules.
The acquisition model (found functioning parts are the current default).
The final content band.
Exact ship functions and departure/credits behavior: exact departure, irreversibility, revival limit, and credits.
Campaign duration.
The leading story/corruption direction.
Oxygen/capacity tuning.
The second-language utterance register remains a relationship to settle; it is not a third guaranteed language.
Remaining detailed regional work: placed objective variants, access constraints, state combinations, rewards, and cross-order checks. Production Scope owns workload estimates.