**Owner-set premise, 2026-09-07: the astronaut and the unofficial space station.** This is the game's premise, not another team candidate. It supersedes incompatible earlier visions and PM provisional selections. Details not specified by the owner remain design proposals. A premise decision is not approval of the complete design or permission to develop.

You are an astronaut. Your ship is broken, forcing you to land at the nearest space station you can find. The station is unofficial. Its emergency signal says **“stay away!”**, and it refuses to answer your hails. With no alternative, you land at its dock anyway.

1. **Limited oxygen away from the ship.** The station's artificial atmosphere is broken. You can carry only a limited amount of oxygen when leaving the ship, and can find additional oxygen canisters around the station. The ship can produce an unlimited amount of oxygen. Running out of oxygen kills you. Unlimited production aboard the ship does not mean unlimited portable oxygen.

Hold the Flood remains a **survivorlike centred on exploration, experimentation and mystery**. The owner's desired reactions remain, in order: “I discovered something incredible,” then “I made a ridiculous build work.” Horde survival and player-developed power must be convincing parts of ordinary expeditions, not merely automatic attacks attached to a sequence of puzzle rooms.

Discovery can reinterpret something visible all along, reveal an unexpectedly larger world, or uncover history and mythology while the player seeks useful power. A notebook full of observations is welcome. Not every historical discovery must unlock a door, upgrade or mechanical action: changed understanding is a valid payoff.

Equipment competing with salvage space remains a point of enthusiasm. Exact rack dimensions, oxygen storage and starting equipment are not fixed by the owner. Enemy manipulation can support a build but is not the central science. An extraordinary combination should result from player choices and learned interactions, not solely permission to trigger one specially authored room.

The ship's rescued people should give exploration an emotional and moral purpose as well as opportunities to improve capabilities. The owner's standing preference is an authored, non-procedural cast the player knows and cares about. No particular cast size, survivor job system, escort mechanic, recurring upkeep or permanent survivor death is approved by the premise.

Clear immediate goals and readable survival information are essential. The player may not understand a language or the revival mystery; they must still understand oxygen remaining, health danger, basic controls and what they are currently trying to achieve. Bad UX is not a mystery mechanic.

- **Expedition pressure:** oxygen capacity/depletion/refill behaviour, carrying cost, canister placement and how retreat, exploration and growing combat power interact. Do not introduce rationing of the ship's unlimited oxygen production by accident.

The owner's ambition remains about **30 hours to apparent credits**, **90 hours total to the real ending**, and **100–200 hours total for the deepest optional secrets**. Those are desired milestones, not measured duration, proven feasibility or permission to fill time with oxygen refills, grinding or repeated solved routes. No district/deck count is fixed.

Unity and Steam are selected. **Design only** until the owner explicitly approves a compelling design and authorizes development: no game code, builds, prototypes, implementation, simulations or playtests. Prefer suitable free existing assets. Purchases and rare authored game assets require owner approval. **No AI-generated art**, including concepts and placeholders.

The harvest-dependence twist and information-hiding songs remain rejected. The drowned city, Hearth/furnace story, Bell House, Choir network and prior neutral clinic expedition are not the canonical setting or progression. Individual interactions or design lessons may be proposed again if they genuinely fit this station; renaming old machinery into space is not an integration plan.

The Visionary shapes and protects this owner-set identity; the Lead integrates the design. Existing specialists develop their domains against these facts. [Tasks](/Hold%20the%20Flood/Tasks) assigns one connected first-sortie design: oxygen, horde/build choices, a discoverable clue, a survivor/rescue opportunity and return/revival consequences, followed by a reason to venture out differently. This is a written design, not a prototype. [Team](/Hold%20the%20Flood/Team) defines responsibilities. Discussion stays in [project chat](/chat/); current rules and consequential unresolved choices stay in articles.

Earlier team-authored proposals, the prior inspiration analysis and numbered decision log remain intact in [the preceding Brief revision](/Hold%20the%20Flood/Brief?revision=e2e390). They are historical evidence, not competing current instructions. Continuing owner constraints are restated above; where an older decision conflicts with this premise, the owner's current premise wins. Historical reviews remain in [PM Reviews](/Hold%20the%20Flood/PM%20Reviews).