**Status.** First-pass planning scaffold, not owner-approved canon and not a promise of duration or production scope. The opening direction is selected; later mechanisms, story interpretations and ending actions below remain proposals. “Apparent,” “deeper” and “secret” describe layers of discovery, not confirmed endings. In particular, irreversible departure and revival being limited to the station field are optional story/gameplay assumptions, not established facts. The 30/90/100–200 hour figures are owner ambitions that this desk outline cannot validate.

| **Opening: one part, two machines** | At Cargo Turn, remove the isolation bridge that the ship needs and bring it aboard. A loose Fork Relay found there may be deployed so the service rack and patient-bay shutter remain powered, costing time and leaving the part installed; keeping it instead preserves the part but consumes carrying space. | Read trigger-to-payload causality, decide whether to deploy the loose part or keep it, survive the resulting shutter state, and learn which property returns after death. | Station infrastructure and combat equipment obey related rules; a “key” can also be a build component. This is the selected Option A direction, while exact rooms, timing and starting rig remain proposed. |

| **Optional secret layers** | Complete personal investigations, rare utterance interactions and the late whole-station connection puzzle. | Connect every actually placed playable room or bound group while preserving every selected local binding, test cross-language hypotheses, and revisit places for consequences rather than checklist residue. The exact graph and joining mechanism remain proposals. | The station itself becomes a manipulable historical object. The deepest discoveries may complicate every prior account without supplying a single omniscient answer. Whether completion freezes the whole layout, and whether any secret is required for a “true ending,” are unresolved. |

Late expedition goals are represented by the deeper and optional-secret rows rather than a separate act: they turn accumulated regional work into central access, consequential action and the entire-station objective.

This is a replaceable middle-game example, not established fiction. After Yeva is home, she recognizes the identifier on a photographed R-3 pressure log but cannot infer where its line now runs. The player seats a relay-family part in Dorn’s collar receiver, listens, retrieves the part, then prepares the expedition. This cooperative preparation costs time aboard; it does **not** create scarcity, consumption, upkeep or a choice between the reading and carrying the relay. Listening yields one pressure value and a suspect channel—not a route or proof.

Inside the station, builder service marks let the player predict which of two valves performs “sample” and which performs “purge.” A visitor-language source/destination pair then predicts that the line was redirected through a decontamination ward. The player tests the prediction in world time, holds a horde at the valve, and—under the replaceable found-parts progression default—finds an **Impact Dynamo** powering the abandoned sampling loop. Removing it ends that shortcut but earns a new build change: motion or impact can now trigger the player’s Barrier Plate or Pulse Projector, enabling a mobile counterwall instead of the opening’s fixed sensor rig.

| Late/deep | As a candidate analogy only, local build grammar may inform station-connection work without making one universal loadout. | Survivors react to central choices according to what they know and what the player shared. | Cross-language evidence opens or reframes a route; utterance interaction remains bounded and never becomes an all-purpose oracle. |

These are comparable scoping proposals, not commitments, production forecasts or implied playtime. A **region** is a coherent expedition territory. A reusable **room module** is an authored layout/encounter pattern; one module may produce multiple **placed playable room instances**, and the late whole-station objective concerns those actual instances, not the catalogue count. A **discovery chain** is a multi-step investigation with a gameplay consequence. A **major set-piece** is an exceptional encounter or sequence and may also host a chain, room or objective; cross-listed units are not additive workload.

| Regions / reusable room modules / placed playable room instances | 4 / 24 / 40 | 7 / 45 / 80 | 10 / 70 / 130 |

| Authored scenes per survivor, including rescue and home follow-through | 3 | 5 | 7 |

| Trap or station-mechanism families / placed instances across both occupant eras | 5 / 12 | 9 / 24 | 14 / 40 |

| Persistent-connection opportunities, including the final station-scale use; mechanism TBD | 2 | 4 | 6 |

Even the STRETCH column does not justify 30, 90 or 100–200 hours. Counts still missing include oxygen canister and carrying-capacity variants, starting loadout alternatives, resource sinks, objective reuse, branch scenes, clue permutations and technical state combinations. Measurements still missing include average expedition and revisit length; how often death creates a genuinely new route rather than repetition; clue discovery and correction rates; viable-build count after dominance testing; horde encounter variety; dialogue/state volume per survivor; gating and backtracking costs; accessibility pacing; and the proportion of players who pursue secrets. Only later implementation, play evidence and production costing can relate inventory to duration or feasibility.

The PM must assign ownership; the roles below are recommendations. Unless labelled “rules only,” each range covers a **SMALL-band specification**, not all authored content in that band. Author rows estimate focused author time and exclude waiting, independent review, implementation, assets and playtesting; the Reviewer row estimates review itself. Each row produces only its named specification and cross-references its predecessor rather than rewriting that predecessor’s rules. Ranges overlap in calendar time and are planning guesses, not proven throughput, deadlines or a complete schedule.

3. Should corruption be accessed mainly through voluntary treatment, or should death itself alter the astronaut? The replaceable default assumes voluntary access and adds no ordinary death penalty; it does not rule out other owner-chosen personal consequences.

5. Is voluntary departure the apparent-credit action; is it irreversible; does leaving end revival; and may the player postpone it indefinitely? The outline uses yes as one replaceable package, not as fact.

7. Can rescued people be lost after reaching the ship, and if so under what visible, non-arbitrary rule? The current outline assumes no answer.