Closed Items Register

This file is append-only history for items that used to be open. Never delete or renumber a row. When an open item closes, move its row here and add how it resolved.

ID	Area	Item	Intent steer	Detail lives in	Blocked by	Priority	Raised	Resolved
CI-001	art	Get the real camera numbers.	-	06_Design Session Decision Log -> Camera	-	DO FIRST	2026-08-25	Done 2026-08-26. Read from src/world/CameraRig.gd and scenes/main.tscn: perspective, 75 degree FOV, fixed 50 degree pitch, zoom 5-40 default 20, yaw orbit exists.
CI-002	docs	Save 06_Design Session Decision Log.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/04_Project Management/06_Design Session Decision Log.md.
CI-003	docs	Save PC Action List.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done, then superseded by docs/open-items.md and deleted during consolidation.
CI-004	docs	Save 07_Art Direction Style Bible.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/03_Art & Audio/07_Art Direction Style Bible.md.
CI-005	docs	Save 23_Port System Design.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/01_Design Documents/23_Port System Design.md.
CI-006	docs	Save 24_Module System Design.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/01_Design Documents/24_Module System Design.md.
CI-007	docs	Save 25_Power System Design.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/01_Design Documents/25_Power System Design.md.
CI-008	docs	Save 26_Pollution, Waste & Organic Systems Design.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/01_Design Documents/26_Pollution, Waste & Organic Systems Design.md.
CI-009	docs	Save 28_Restoration - Third Endgame Path.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/01_Design Documents/28_Restoration - Third Endgame Path.md.
CI-010	docs	Save 03_Data Pipeline & Design Console - Plan.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/02_Technical Documents/03_Data Pipeline & Design Console - Plan.md.
CI-011	docs	Save 00_Canonical Design Facts.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/04_Project Management/00_Canonical Design Facts.md.
CI-012	docs	Save 20_Belt Logistics - The Router.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/01_Design Documents/20_Belt Logistics - The Router.md.
CI-013	docs	Save 22_Trains & Rail.md to the vault.	-	PC Action List -> B	-	HIGH	2026-08-25	Done. File lives at docs/01_Design Documents/22_Trains & Rail.md.
CI-014	docs	Save 03_Idea - Network, IoT & Factory Automation.md to the vault.	-	PC Action List -> B	-	NORMAL	2026-08-25	Done. File lives at docs/00_Ideas/03_Idea - Network, IoT & Factory Automation.md.
CI-015	docs	Save 04_Idea - Quantum Silo.md to the vault.	-	PC Action List -> B	-	NORMAL	2026-08-25	Done. File lives at docs/00_Ideas/04_Idea - Quantum Silo.md.
CI-016	docs	Save base_chevron_explainer.html to the vault.	-	PC Action List -> B	-	NORMAL	2026-08-25	Done. File lives at docs/05_Reference/base_chevron_explainer.html.
CI-017	docs	Reconcile 04_Factory module section to fixed roof slots.	-	PC Action List -> C1	-	HIGH	2026-08-25	Done 2026-08-26.
CI-018	docs	Reinstate power poles in 04_Factory.	-	PC Action List -> C1	-	HIGH	2026-08-25	Done 2026-08-26.
CI-019	docs	Relabel configurable connection range as backend tuning.	-	PC Action List -> C1	-	HIGH	2026-08-25	Done 2026-08-26.
CI-020	docs	Reclassify Chemical Gas as a fluid in design.	transport medium decides	PC Action List -> C1	-	HIGH	2026-08-25	Done 2026-08-26. Data intentionally left unchanged pending fluid-input path.
CI-021	logistics	Collapse Splitter, Merger, and Filter Splitter into Router in design.	-	PC Action List -> C1	-	HIGH	2026-08-25	Done 2026-08-26.
CI-022	storage	Document the shipped ChestMode / future StorageRole system.	-	PC Action List -> C1	-	HIGH	2026-08-25	Done 2026-08-26 in 04_Factory.
CI-023	docs	Extract shared tables from authority documents instead of copying them.	-	PC Action List -> C1	-	NORMAL	2026-08-25	Done 2026-08-26. Five tables verified byte-identical.
CI-024	docs	Rewrite CLAUDE.md / root AI wrapper state against current code.	-	PC Action List -> C6	-	HIGH	2026-08-25	Done 2026-08-26. Root issue was AGENTS.md; wrappers are thin pointers again.
CI-025	tooling	Fix npm run check:ai and npm run check:docs not running on the PC.	-	PC Action List -> C6	-	HIGH	2026-08-26	Done 2026-08-26 by installing PowerShell 7.6.5.
CI-026	vertical	Design vertical power routing across levels.	-	PC Action List -> G	-	HIGH	2026-08-25	Done. See Vertical Transport & Cross-Level Power.md.
CI-027	art	Lock exact palette hex values.	-	06_Design Session Decision Log -> Open Questions	-	HIGH	2026-08-25	Resolved in 07_Art Direction Style Bible section 1.
CI-028	art	Restate grid unit to Blender unit ratio for the new style.	-	06_Design Session Decision Log -> Open Questions	-	HIGH	2026-08-25	Resolved in 05_Asset List & Delivery Specification: 1 Godot unit = 1 grid cell = 1 metre.
CI-029	art	Decide category accent colours versus locked palette.	-	06_Design Session Decision Log -> Open Questions	-	HIGH	2026-08-25	Resolved. Category colour-coding dropped; buildings are distinguished by silhouette and Accent Orange is the only accent.
CI-030	storage	Decide whether to collapse two solid storage buildings.	-	06_Design Session Decision Log -> Open Questions	-	HIGH	2026-08-25	Resolved. One Silo, capacity by Mk.
CI-031	power	Decide whether Mk levels increase power draw.	-	25_Power System Design -> Mk Levels and Power Draw	-	HIGH	2026-08-25	Resolved. Yes, both Base Mk and Machine Mk increase power draw.
CI-032	power	Decide above-ground / underground / belt interaction with power routing.	-	25_Power System Design -> Open Questions	-	HIGH	2026-08-25	Resolved in Vertical Transport & Cross-Level Power.md: one shaft per type, conduits are rated links, transmission horizontal-only.
CI-033	pollution	Decide whether pollution travels between vertical levels.	-	26_Pollution, Waste & Organic Systems Design -> Pollution and Vertical Levels	-	HIGH	2026-08-25	Resolved. Air does not travel between levels; ground contamination leaches downward.
CI-034	pollution	Decide pollution HUD readout.	-	26_Pollution, Waste & Organic Systems Design -> Open Questions	-	NORMAL	2026-08-25	Resolved. Pollution tab specced in 17_HUD Panel Flap.
CI-035	power	Decide whether transformers use the standard base/port system.	-	25_Power System Design -> Transformers Use the Standard Base	-	NORMAL	2026-08-25	Resolved. Transformers use the standard base with all ports disabled.
CI-036	power	Decide transmission line blocking model.	-	25_Power System Design -> Transmission Line Clearance	-	NORMAL	2026-08-25	Resolved. Height-based clearance for horizontal transmission.
CI-037	power	Decide whether Outpost Mode and main game share power code.	-	25_Power System Design -> Shared Power Code with Outpost Mode	-	NORMAL	2026-08-25	Resolved. Share the solver, not content.
CI-038	logistics	Decide what happens when a filtered Router output backs up.	priority is overflow	20_Belt Logistics - The Router -> Priority IS the overflow mechanic	-	NORMAL	2026-08-25	Resolved. Backed-up output loses its turn and items waterfall to the next priority.
CI-039	logistics	Decide whether the Incinerator recovers anything.	capped below embodied energy	20_Belt Logistics - The Router -> The deliberate way out	-	NORMAL	2026-08-25	Resolved. It recovers heat as power, never profitably.
CI-040	trains	Decide whether trains require Fuel Cells or abstract locomotive power.	-	22_Trains & Rail -> Related Existing Question	-	NORMAL	2026-08-25	Resolved. Neither; fuel type distinguishes locomotive types.
CI-041	research	Resolve Research Tree open questions as of v0.3.	-	08_Research Tree Design -> Open Questions	-	NORMAL	2026-02-28	Resolved in v0.3 revision history.
CI-042	ui	Decide Tutorial/onboarding UI.	-	09_UI-UX Design Document -> Open Questions	-	NORMAL	2026-03-31	Resolved. Guided step-by-step tooltips overlaid on real session, skippable, Genesis as guide voice.
CI-043	art	Decide Genesis visual design in the old art document.	flying bot, compact head, visor laser, no legs	01_Art Direction Document -> Open Questions	-	NORMAL	2026-02-22	Resolved in the old art document; later canonical status is tracked separately in OI-129.
CI-044	art	Decide whether crashed ship wreckage is fixed or procedural.	fixed asset	01_Art Direction Document -> Open Questions	-	NORMAL	2026-02-22	Resolved. Fixed asset.
CI-045	art	Decide how rail tracks look in the old low-poly style.	slightly raised above terrain	01_Art Direction Document -> Open Questions	-	NORMAL	2026-02-22	Resolved. Slightly raised above terrain.
CI-046	logistics	Decide whether belts are flat or raised.	slightly raised	01_Art Direction Document -> Open Questions	-	NORMAL	2026-02-22	Resolved. Slightly raised above terrain.
CI-047	art	Decide whether day/night affects building glow intensity.	yes	01_Art Direction Document -> Open Questions	-	NORMAL	2026-02-22	Resolved. Emissive glow increases at night and dims during day in the old design.
CI-048	world	Decide whether planets have names.	procedurally generated by default	05_World Generation & Planet Types -> Open Questions	-	NORMAL	2026-02-26	Resolved. Planets have procedural names by default.
CI-049	world	Decide whether players can rename planets.	yes	05_World Generation & Planet Types -> Open Questions	-	NORMAL	2026-02-26	Resolved. Any landed planet can be renamed for the session.
CI-050	world	Decide starting planet naming.	fixed lore name	05_World Generation & Planet Types -> Open Questions	-	NORMAL	2026-02-26	Resolved. Fixed lore name, not procedural; exact name remains in World Bible scope.
CI-051	outpost	Decide whether Outpost co-op uses both Genesis bots or one.	one Genesis	06_Outpost/01_Outpost Mode Overview -> Open Questions	-	NORMAL	2026-03-13	Resolved. One Genesis.
CI-052	outpost	Decide whether enemies attack Genesis.	aliens yes, external drones no, internal drones yes	06_Outpost/01_Outpost Mode Overview -> Open Questions	-	NORMAL	2026-03-13	Resolved. Aliens and internal drones attack Genesis; external drones do not.
CI-053	outpost	Decide physical presence versus remote management panel for Commander role.	physical presence only	06_Outpost/01_Outpost Mode Overview -> Open Questions	-	NORMAL	2026-03-13	Resolved. Physical presence only.
OI-024	research	Rewrite Research Tree Branch 8 for Module Technology, fixed slots, and no stacking.	-	08_Research Tree Design -> Branch 8; 24_Module System Design -> Research node names	-	HIGH	2026-08-25
OI-016	pollution	Decide whether waste systems get their own research branch or join an existing branch.	-	26_Pollution, Waste & Organic Systems Design -> Waste Handling - Three Choices	OI-242	HIGH	2026-08-26	Decided 2026-08-27. Branch 9, Environmental - waste treatment, remediation and farming together. Building count was the weakest argument; discoverability of an endgame path was the strongest. Restoration is canonical as one of three endgame routes (Canonical section 12), and a route whose toolset is scattered across branches named for other things is not a route. The three systems are coupled by real dependencies, not theme: farms feed Soil Remediation, farms feed the Biomass Burner, recycled organic waste feeds farms as fertiliser. It also has a genuine Tier 1 anchor in the Biomass Burner, so the branch is not a locked region the player can see but not enter. Implementation is OI-248, blocked on OI-247.
OI-012	docs	task	Merge Vertical Transport & Cross-Level Power.md into 06_Vertical Layers & Core Power System.md, then remove the root-level addendum.	-	-	-	HIGH	2026-08-25
OI-014	code	task	Replace the PLACE_SHAFT Genesis task with type-specific vertical placement tasks.	-	06_Vertical Layers -> Shafts	OI-013	HIGH	2026-08-25
OI-015	power	undecided	Resolve Solar Panel vs Solar Collector naming.	-	06_Vertical Layers -> Open Questions	-	NORMAL	2026-08-25
OI-034	docs	task	Add Restoration as a third endgame path in 01_Master GDD.	-	28_Restoration - Third Endgame Path; Pollution -> Restoration as a third endgame path	-	HIGH	2026-08-25
OI-002	ports	task	Check git history for why the pre-Feb 2026 port/connector-slot system was removed.	-	23_Port System Design -> Inherited Constraints	-	DO FIRST	2026-08-25
OI-268	farming	task	Correct Biomass Burner fuel from "Wood (solid belt)" to raw_biomass.	-	25_Power System Design -> Source Table; 04_Factory -> Power Buildings	-	HIGH	2026-08-27
OI-269	farming	task	Rewrite Pollution section 7: three crops become two families.	-	26_Pollution, Waste & Organic Systems Design -> Organic Farming	-	HIGH	2026-08-27
OI-247	design	defect	25_Power System Design and 26_Pollution, Waste & Organic Systems Design both assert an organic production chain that does not exist. No resources, no buildings, no recipes. Pollution section 7 lists fast-growth timber, algae and bio-polymer crops as farm outputs; none is defined (raw_wood appears in zero files, algae and biofuel in none). No farm building exists anywhere in the vault - only building_biomass_harvester, which extracts wild biomass, a different thing from cultivated crops. Consequences today: the Biomass Burner is a generator with undefined fuel (25_Power System Design section 5.1 gives it "Wood (solid belt)"), and Soil Remediation consumes an organic input nothing produces. This is the same "already specced" failure recorded against Geothermal Tap and Core Power Station - a document asserting something exists is what stops anyone checking. Needs a design pass that invents the chain, not a reconciliation pass; wants a conversation rather than a prompt.	-	26_Pollution, Waste & Organic Systems Design -> Organic Farming; 25_Power System Design -> Source Table; 04_Factory -> Power Buildings	-	DO FIRST	2026-08-27
OI-062	logistics	undecided	Decide whether a Fluid Router exists or pipes junction implicitly.	-	20_Belt Logistics - The Router -> What Stays Separate	OI-004	NORMAL	2026-08-25
OI-271	data	task	Give Chemical Gas a phase, in three steps that must move together. A Router is configured ITEM / LIQUID / GAS and a gas Router cannot route liquid, but nothing in the code says which fluids are gases. (1) Add enum Phase { LIQUID, GAS } and @export var phase to FluidDefinition — not a new ResourceTier value, which would turn every tier == FLUID check into a two-way test, contradict Canonical section 3, and renumber an enum serialised as a raw int into .tres. (2) Create data/fluids/raw_chemical_gas.tres as a FluidDefinition and set phase = GAS; it does not exist today, so Chemical Gas has no viscosity, density or temperatures and nowhere to put a phase. (3) Retype data/resources/raw_chemical_gas.tres to tier = 4 (OI-010). Any one alone leaves an inconsistent resource.	enum on FluidDefinition, never a new tier	06_Design Session Decision Log -> Transport & Logistics; FluidDefinition.gd; ResourceDefinition.gd	OI-010	HIGH	2026-08-27
OI-131	logistics	undecided	Design drones/delivery bots beyond the current four bullets. Settled 2026-08-27: fluid only via barrels (OI-272) or a fluid-carrier variant; battery-limited with a recharge cycle; drones draw power by hooking onto power poles, giving the distribution tier a second job; drones cannot fly underground and change level through a vertical shaft, so underground logistics stays belts, pipes and rail. Still undesigned: payload, range, speed, Drone Port behaviour, routing, fleet rules, and whether power cost is per-trip or port-side.	Drone Port, routes, payload, charging	06_Design Session Decision Log -> Transport & Logistics	-	DEFERRED	2026-08-25
OI-312	docs	task	State the lossy-loop guard rail once, as a principle, and reference it from each system. It now appears four times independently: Incinerator heat recovery, the flue capture ceiling, nuclear reprocessing, and box recycling. The principle is that a loop which closes completely removes the decision it was meant to create. Four occurrences is a pattern, and restating it per system invites one of them being written without it. Natural home is 00_Canonical Design Facts.	one statement, referenced from each system	06_Design Session Decision Log -> Cache Boxes	-	HIGH	2026-08-29
OI-275	logistics	task	Write a Transport & Logistics design document once the carrier discussion completes. Decisions currently live only in the decision log: belts as queues versus pipes as pools, the Router's three configurations, drone barrels/charging/underground exclusion, and quantum travel's scope. Carriers still to cover when the session resumes.	-	06_Design Session Decision Log -> Transport & Logistics	-	HIGH	2026-08-27
OI-041	tooling	undecided	Decide the doc naming rule and make npm run check:docs match it. Now reports 21 findings, and at least two are false positives: open-items.md and closed-items.md are flagged as root-level strays, but they live at docs/ root deliberately and AGENTS.md points at that exact path. Of the 18 NN_ prefix findings, 10 are files cited by exact path in the 04_Factory Design Document Index, so renaming them would break the index. The likely answer remains narrowing the check, not renaming the docs. Genuine strays: ROADMAP.md and ROADMAP-COMPLETED.md.	narrow the check, do not rename	AGENTS.md -> Design Document Index	-	HIGH	2026-08-26
OI-007	docs	task	Fix 01_Master GDD camera wording; orbit is implemented, not pan-and-zoom only.	-	-	-	HIGH	2026-08-26
OI-011	docs	task	Merge HUD Pollution Tab Specification.md into 17_HUD Panel Flap.md, then remove the root-level addendum.	-	-	-	HIGH	2026-08-25
OI-013	docs	defect	Replace the generic Shaft model with Item Lift, Fluid Riser and Power Conduit in the remaining documents. 06_Vertical Layers is done (2026-08-27); 01_Master GDD, 08_Research Tree Design, 03_Procedural Generation Reference and 06_Design Session Decision Log still reference the generic Shaft. The Decision Log is append-only and should be left alone.	-	06_Vertical Layers -> Vertical Structures	-	NORMAL	2026-08-25
OI-021	docs	defect	Remove the obsolete Global Factory Overview F section and related tutorial references.	-	-	-	HIGH	2026-08-26
OI-022	docs	defect	Remove stale emissive level ramp references from 24_Module System Design.	-	-	-	HIGH	2026-08-26
OI-005	power	task	Define brownout shedding hysteresis, staged shedding, shutdown threshold, and recovery threshold.	recovery threshold strictly above shutdown threshold	25_Power System Design -> Implementation Trap - Brownout Oscillation	-	DO FIRST	2026-08-25
OI-079	power	undecided	Decide all PowerTuning loss and range placeholder values.	-	25_Power System Design -> Backend Tuning Values	-	HIGH	2026-08-25
OI-082	power	undecided	Decide voltage drop coefficient and load-to-reach curve shape.	-	25_Power System Design -> Open Questions	-	HIGH	2026-08-25
OI-086	power	undecided	Decide whether under-voltage is its own status state or shares amber with stalled.	-	25_Power System Design -> Brownout reads on the base rings	-	NORMAL	2026-08-25
OI-101	power	undecided	Decide Incinerator and Air Scrubber power draw.	-	04_Factory -> Waste & Environmental Buildings	-	HIGH	2026-08-26
OI-104	power	undecided	Decide generator output values for Biomass Burner, Nuclear Reactor, Geothermal Tap, and Core Power Station.	very low; high; moderate-high; very high	04_Factory -> Power Buildings	OI-001 for Nuclear Reactor	HIGH	2026-08-26
OI-003	ports	undecided	Decide direct building-to-building transfer between touching ports.	allow at reduced rate	23_Port System Design -> Direct Building-to-Building Transfer	-	DO FIRST	2026-08-25
OI-060	ports	undecided	Decide Base Mk throughput multipliers.	-	23_Port System Design -> Base Mk - Port Upgrades	-	NORMAL	2026-08-25
OI-065	logistics	undecided	Decide whether a Router can feed another Router directly port-to-port.	-	20_Belt Logistics - The Router -> Open Questions	-	NORMAL	2026-08-25
OI-058	ports	undecided	Decide default port state: all disabled or auto-enable on adjacent belt.	deferred to playtest	23_Port System Design -> Port configuration	-	NORMAL	2026-08-25
OI-259	ports	undecided	No named mechanism for a player reconfiguring a port. 23_Port System Design section 9 removes PLACE_CONNECTOR and PLACE_PIPE_CONNECTOR and says ports come into existence with their building, but never says how a player changes a port from DISABLED to OUTPUT, or sets a filter. Either it is an instant UI action - which collides with the AGENTS.md rule that all placement is queued as a Genesis task, never instant - or it is a new task type such as CONFIGURE_PORT. This decides whether GenesisTask.Type shrinks by two or shrinks by two and grows by one, so it blocks the refactor.	-	23_Port System Design -> Genesis Task Types; GenesisTask.gd:11-21	-	HIGH	2026-08-27
OI-258	testing	task	Write end-to-end transport scenarios BEFORE the port refactor, against the current system. There is no GUT test and no ScenarioTestRunner scenario anywhere that exercises producer -> connector -> belt -> connector -> consumer. The three existing scenarios are all Genesis fetch/carry/build. The refactor replaces the item-transport mechanism with zero regression coverage on the behaviour being replaced, so scenarios must be written first to be runnable before and after.	write against the current system first	ScenarioTestRunner.gd:93-105; AGENTS.md -> Testing Scope	-	DO FIRST	2026-08-27
OI-320	ui	task	Recipe selection is not designed anywhere. accepted_recipe_ids is multi-valued in shipping data - Assembly Machine and Advanced Assembly carry three recipes each, Press Machine and Advanced Press two - but no document says how a player picks one. Nothing in 04_Factory or 09_UI-UX Design Document covers it. It is a prerequisite, not a nicety: 23_Port System Design §5.1b makes recipe selection step 2 of three, and §5.1c gates inbound port auto-enable on it, so a building with no recipe accepts nothing. Needs: where the choice lives (Building Detail Panel, beside port config), whether a single-recipe building shows the control at all, how extraction buildings present in the same panel, and what a recipe change does - to a running job, to buffered inputs, to auto-enabled ports (they follow the new recipe, their filter being empty) and to explicitly filtered ports whose filter the new recipe cannot use, which should be surfaced rather than silently cleared since the player set them deliberately.	panel owns recipe and ports together; never silently clear a player's filter	23_Port System Design -> §5.1b, §5.1c; 04_Factory -> Recipes; BuildingDefinition.gd	-	HIGH	2026-08-30
OI-020	storage	undecided	Resolve the two-solid-storage-building conflict; canonical design has one Silo.	-	AGENTS.md -> Known Code/Design Divergences; 04_Factory -> Storage Buildings	-	HIGH	2026-08-26
OI-059	ports	undecided	Verify one port per face is enough throughput for the 1x1 Silo.	-	23_Port System Design -> Port count	OI-020	HIGH	2026-08-25
OI-063	logistics	undecided	Decide whether the Smart Buffer remains distinct now that ports filter.	-	20_Belt Logistics - The Router -> Open Questions	OI-020	HIGH	2026-08-25
OI-068	storage	undecided	Decide Silo capacity curve.	suggested 500 / 1000 / 2000	-	OI-020	NORMAL	2026-08-25
OI-069	storage	undecided	Decide whether Mk raises how many distinct resource types a Silo can hold.	would make Mk3 Silo a sorting hub	06_Design Session Decision Log -> Open Questions	OI-020	NORMAL	2026-08-25
OI-242	research	defect	The research tree was never updated for any of the design work, and may describe a system that no longer exists. Not four separate gaps - one fact with four consequences (OI-243, OI-006/OI-016, OI-244, OI-245). Scoped individually they look like four small gaps; scoped together, research is the least reconciled system in the vault. Decide whether the tree is amended in place or rebuilt against current design.	-	08_Research Tree Design; 25_Power System Design -> The Three Tiers; 26_Pollution, Waste & Organic Systems Design -> Waste Handling - Three Choices	-	DO FIRST	2026-08-27
OI-006	research	undecided	Decide the research unlock path for waste, environmental, and new generator buildings.	-	26_Pollution, Waste & Organic Systems Design -> Waste Handling - Three Choices	OI-242; OI-248	DO FIRST	2026-08-26
OI-243	research	defect	04_Factory and 08_Research Tree Design describe two different research trees under the same name. Nine of ten node names taken from 04_Factory appear nowhere in 08: Basic Logistics, Ore Processing, Basic Power, Advanced Logistics, Fluid Handling, Filter Logistics, Power Management, Rail Transport, Module Technology. Only Chemical Processing appears in both. Every research reference in the vault is therefore ambiguous about which tree it points at, including the Module Technology nodes written during the 04_Factory pass. Fix: 04_Factory's tree becomes a pointer, and the nine orphan names get audited - real design in 08 under different names, or genuinely absent. Run as its own pass once decided.	08 is the authority: it is the dedicated document, and 04_Factory calls its own version a summary	04_Factory -> Research System; 08_Research Tree Design -> Branch Overview	-	DO FIRST	2026-08-27
OI-025	research	task	Check Power research branch against the three-tier power model.	-	25_Power System Design -> Open Questions	-	HIGH	2026-08-25
OI-087	power	undecided	Decide whether transmission needs its own research branch or extends Branch 2.	-	25_Power System Design -> Open Questions	OI-025	NORMAL	2026-08-25
OI-244	research	defect	Transmission has no research nodes. Canonical section 8 has Generation -> Transmission -> Distribution. Branch 2 unlocks Power Conduit and Power Pylon under Power Grid Expansion, but there are no transformer nodes, no transmission lines and no voltage tiers - and transmission is the entire research path in 25_Power System Design section 3.2. The tree has no way to reach it. Distribution correctly needs nothing, since poles are baseline. Adding nodes means inventing RP costs.	-	25_Power System Design -> The Three Tiers; 08_Research Tree Design -> Branch 2	-	HIGH	2026-08-27
OI-245	research	defect	Biomass Burner, Wind Turbine and Nuclear Reactor have no research nodes. Branch 2 covers Fuel, Solar, Core Power and Geothermal only. All three are specified in 25_Power System Design section 5.1 and were added to 04_Factory in the ten-buildings pass. Adding nodes means inventing RP costs.	-	25_Power System Design -> Source Table; 08_Research Tree Design -> Branch 2	-	HIGH	2026-08-27
OI-246	research	task	Audit the remaining 08_Research Tree Design branches (1, 3, 4, 5, 6, 7) against current design. Branch 8 and Branch 2 both proved stale, so the others should not be assumed current.	-	08_Research Tree Design	-	NORMAL	2026-08-27
OI-248	research	task	Create Branch 9, Environmental - waste treatment, remediation and farming. Decision made (see closed OI-016); this is the implementation. Scope it as one pass that defines the organic chain and its research together, because nodes for buildings and resources that do not exist would just move the hole. Covers the five environmental buildings, the farm chain, and Restoration as a visible endgame route.	Environmental, not Waste - farming belongs in it	08_Research Tree Design -> Branch Overview; 26_Pollution, Waste & Organic Systems Design	-	HIGH	2026-08-27
OI-322	research	task	Two node families are referenced by other documents but exist in no tree. Power Management gates the Power Priority System (04_Factory -> Power Priority System), and Bot Programming I/II gate the Construction Bot Bay and Genesis automation (01_Master GDD, and three places in 04_Factory). Both were defined only in the tier-gated 04_Factory listing deleted under OI-242, and because that listing contradicted the tree's structure they were never really in the tree at all - the fork made it look as though they were. This is the same class of gap as OI-244 and OI-245, and should be done with them rather than separately, since all three mean adding nodes and inventing RP costs in one balancing pass. Power Management belongs in Branch 2, Bot Programming in Branch 7 (Genesis Abilities) or Branch 1. Do not delete the four references - they state real design intent; it is the nodes that are missing, not the claims that are wrong.	add the nodes, keep the references	08_Research Tree Design -> §0.2; 04_Factory -> Power Priority System; 01_Master GDD	-	HIGH	2026-08-30
OI-081	research	undecided	Verify no win-critical recipe sits behind an optional research branch.	must not	25_Power System Design -> Mk is a build requirement, not a research node	-	HIGH	2026-08-25
OI-276	data	task	Replace ResourceTier with category, substance, type and phase on ResourceDefinition. type (RAW/REFINED/COMPONENT) is what tier was reaching for; the word and the mixed-in FLUID value were wrong, not the instinct. The tier enum conflates production depth (RAW/REFINED/COMPONENT/ADVANCED) with transport medium (FLUID), so a resource that is raw by origin and fluid by medium cannot be described correctly. Drop tier entirely. It is not serialised, it is read once as a fallback in FactoryStats whose primary path already derives category from the producing building, and that reader bypasses it for fluids with a begins_with("fld_") check. Production depth is derivable from the recipe graph. Transport is derived from phase, not stored - solids ride belts, liquids and gases ride pipes, as a rule in code rather than a field. 19 .tres files. Do this before the port refactor (OI-004): ports gating on phase is cleaner than ports gating on an enum about to be deleted.	four axes; ids stay as stable handles, nothing collapses	06_Design Session Decision Log -> Resource Data Model; ResourceDefinition.gd; FactoryStats.gd	-	DO FIRST	2026-08-28
OI-298	docs	defect	The Drone Port survives in two roadmap documents after being deleted from the design. 01_Development Roadmap.md has "Drone Port building implemented" and "Drones consume power at Drone Port"; ROADMAP.md has "Implement the Drone Port building" and lists Drone Port power consumption under M2.4. Both describe a building that no longer exists - charging moved to the drone module, construction to a dedicated building, fleet visibility to a HUD tab. Note 01_Development Roadmap.md is already ~5 months stale (recorded in AGENTS.md) and ROADMAP.md is a root-level stray flagged for relocation or archiving, so this may be absorbed by fixing those rather than as its own task.	-	06_Design Session Decision Log -> Drones	-	NORMAL	2026-08-28
OI-314	docs	task	Roadmap milestone M3.4 is named for a removed screen. It reads Advanced Diagnostics & Factory Overview, but the full-screen Factory Overview was removed in 09_UI-UX Design Document v0.6 and 05_Reference/04_Data Dictionary lists it as deprecated. The diagnostics half survives - Supply Chain Overlay, Shortage Alerts, the Analytics tab - so this is a rescope, not a deletion. Appears in both 01_Development Roadmap Phase 3 and docs/ROADMAP.md. Same shape as OI-040, and left unfixed for the same reason: renaming a milestone is a roadmap decision, not a wording fix.	rescope, do not silently rename	01_Development Roadmap -> Phase 3; 09_UI-UX Design Document -> Global Factory Overview	-	NORMAL	2026-08-29
OI-036	docs	task	Backfilled 2026-08-30 for the merged roadmap; kept open for the detail. The five-month gap after M2.3 (2026-03-13) is no longer silent: 04_Project Management/ROADMAP.md now carries a Recently completed list reconstructed from git history - the resource schema replacement (2026-08-28), the flap panel system and autoloads (2026-07-05), the HUD components (2026-04-03), the Genesis carry/fetch/build flow (2026-03-31) and the Outpost Mode prototype (2026-03-16) - and a Where we actually are section stating plainly that M2.4 to M2.6 were skipped rather than completed, that M2.7 was built out of order, is roughly two-thirds done and was never marked, and that Outpost Mode sits outside the milestone sequence entirely. That was the dangerous part: the numbering read as linear progress, which is how someone concludes M2.4 is next. What remains is per-milestone detail rather than position: M2.7's completed items have no dates or hours recorded against them, and the Outpost Mode work has no milestone of its own despite being a substantial prototype. Decide whether Outpost gets numbered milestones or stays deliberately outside the sequence.	position is recorded; detail is not	04_Project Management/ROADMAP.md -> Where we actually are; AGENTS.md	-	NORMAL	2026-08-26
OI-324	docs	defect	04_AI Tool Selection Guide.md links to four files that live outside docs/ and therefore cannot resolve on the published site: AGENTS.md, CLAUDE.md, codex.md and .github/copilot-instructions.md. MkDocs resolves them relative to the containing page, producing 04_Project Management/AGENTS.md, which does not exist. Surfaced by the first mkdocs build. Fix by pointing them at GitHub URLs rather than relative paths, since those files are repo-root artefacts and will never live in docs/. Likely absorbed by OI-039, which already calls for that document to be rewritten or retired.	GitHub URLs, not relative paths	04_Project Management/04_AI Tool Selection Guide.md	OI-039	NORMAL	2026-08-30
OI-326	tooling	task	Switch the docs workflow to mkdocs build --strict once OI-313 closes. The gate is deliberately non-strict today: with mkdocs-roamlinks-plugin resolving [[wikilinks]], the 23 links at OI-313's nine missing documents would each fail the build and the site would never publish. ⚠️ Two known false positives will need handling first: [[wikilinks]] and [[Document Name]] appear as examples in prose (03_Data Pipeline & Design Console - Plan), inside backticks, and the plugin does not respect code spans. Escape them or the strict build fails on documentation about wikilinks.	strict only after OI-313 and the two literals	.github/workflows/docs.yml; 03_Data Pipeline & Design Console - Plan	OI-313	NORMAL	2026-08-30
OI-327	docs	defect	Rewrite 05_Reference/08_AI Toolchain.md. Do NOT delete it - unlike 04_AI Tool Selection Guide (OI-039), it is the only place several things exist. Unique and worth keeping: the audio toolchain (Suno and Udio for music, ElevenLabs and Freesound for SFX, Audacity for editing) - audio appears nowhere else in the vault; the licensing checklist of nine tools to verify commercial rights on before shipping; the asset naming convention; the writing and narrative workflow. Stale and contradicting current design: "Apply cel-shader material" and "the cel-shader is the single most important visual system - never skip it"; the Master Style Prompt with thick dark outlines and flat colour faces (OI-249, OI-027); the badge_mesh faction badge band (OI-250); "GitHub Copilot for implementation" and the Claude.ai-designs/Copilot-codes split (same staleness as OI-039); an Obsidian Vault Structure block describing a pre-repo OneDrive folder layout.	rewrite, keep the audio and licensing content	05_Reference/08_AI Toolchain.md	OI-026	HIGH	2026-08-30
OI-328	docs	defect	Found the origin of five of OI-313's nine phantom documents. 08_AI Toolchain -> Obsidian Vault Structure lists Spaceship Design & Building System.md, Combat & War System Design.md, Space Exploration Design.md, Economy & Progression Design.md and Cross-Platform Strategy.md inside a folder tree presented as fact, which is almost certainly why other documents began linking to them. They were planned here in February 2026 and never written. This does not change OI-313's decision - write or delink, never stub - but it does mean the vault structure block must be corrected in the same pass, or the phantom documents are re-legitimised every time someone reads it.	correct the tree, or the phantoms come back	08_AI Toolchain -> Obsidian Vault Structure	OI-313; OI-327	NORMAL	2026-08-30
OI-026	art	task	Rewrite the old Art Direction Document for the current clean chamfered hard-surface style.	-	01_Art Direction Document -> Open Questions; 07_Art Direction Style Bible	-	HIGH	2026-08-25
OI-027	art	defect	Replace the Recraft master style prompt that still specifies the deleted cel-shader direction.	-	01_Art Direction Document -> Recraft Master Style Prompt	OI-026	HIGH	2026-08-25
OI-305	docs	defect	09_UI-UX Design Document specifies the Building Detail Panel and Module UI on six removed decisions. Researched module slots; a whole Module Stack UI section built on stacking; Stack position: 2 of 3; Combined effect: +45%; a Speed Module II implying confirmed module tiers when they are OPEN; and a status line reading "uses building state colours (green/yellow/orange - matches building mesh colour)", which is both the green-for-running model replaced by healthy-is-unlit and the whole-building tint the design forbids. This is the document someone would implement the module screen from.	rewrite against fixed roof slots and the four signal channels	06_Design Session Decision Log -> Module Installation UI; 09_UI-UX Design Document -> Building Detail Panel	-	HIGH	2026-08-29
OI-306	ui	task	Specify module installation and in-place upgrade in the Building Detail Panel. One screen lists roof slots, what is installed and what can go in them, and upgrades a module where it sits rather than crafting a replacement and swapping it. Reuses the Upgrade Kit pattern - crafted item, applied by Genesis in place, changes level and nothing else - which delivers two functions in one slot without reviving stacking. Supersedes the combined-module-crafted-from-both proposal in OI-303.	upgrade in place, mirroring Upgrade Kits	06_Design Session Decision Log -> Module Installation UI; 09_UI-UX -> Building Detail Panel	OI-305	HIGH	2026-08-29
OI-307	ui	task	Surface prerequisites at the point of choice, in the Building Detail Panel and the research tree. A drone module with no packing anywhere is a dead investment and the player finds out only when nothing moves. Generalise it: the panel's job is not only to show state but to show what this building still needs in order to work. The same trap already exists for ports, which default to DISABLED, so a newly placed building does nothing and nothing says why.	show what is missing, not just what is	06_Design Session Decision Log -> Module Installation UI; 23_Port System Design -> Default on placement	OI-305	HIGH	2026-08-29
OI-286	data	task	Split raw_biomass into wood and algae. biomass is a category; wood and algae are substances inside it, so raw_biomass is named one level too high - the equivalent of a resource called metal instead of iron. Touches the Biomass Harvester output, the Biomass Burner's accepted fuel, ref_polymer's inputs, and Soil Remediation's organic input. Partly reverses the farming pass, which was right that three crops feeding identical outputs was duplication but collapsed to the category name rather than a substance name.	wood and algae are substances, biomass is their category	06_Design Session Decision Log -> Resource Properties and Simulation Level; 04_Factory -> Resource List	-	HIGH	2026-08-28
OI-138	tooling	task	Build editing UI for the data console only if DB Browser proves annoying.	low priority	03_Data Pipeline & Design Console - Plan -> Console - Staged, and NOT UI-First	OI-136	DEFERRED	2026-08-25
OI-333	data	defect	smelt_fuel_cell is listed by no building, so it can never run. No .tres in data/buildings/ names it in accepted_recipe_ids, which means the recipe exists, produces ref_fuel_cell, and is unreachable. ref_fuel_cell is in turn an input to assemble_power_cell, so the Power Cell chain is broken at that link - the Chain Explorer renders it as ⚠ no building runs this partway up an otherwise complete tree. Found by the recipe exporter on its first run, which now reverse-maps accepted_recipe_ids and reports any recipe nothing claims. Likely just a missing entry on the Smelter, but check whether Fuel Cells were meant to move to the Chemical Plant, since the recipe takes Chemical Gas.	probably one line on the Smelter	data/recipes/smelt_fuel_cell.tres; data/buildings/smelter.tres	-	HIGH	2026-08-31
CI-054	progression	How Phase 2 reaches players is unresolved, and it is the main risk the two-phase collapse creates. 02_Scope & MVP Document v0.4 is explicit that two design phases is not two releases: Phase 2 is roughly 1,400 to 2,100 hobby-pace hours, which is three and a half to four years of Early Access silence if shipped as one drop. The options are a single 1.0 release, or a sequence of Early Access updates that opens space first and war later, which is the old phase boundary reappearing as a release schedule rather than a design one. That is legitimate, and it needs deciding before the roadmap is rewritten.	belongs to the phasing pass	02_Scope & MVP Document Release Philosophy; ROADMAP	-	HIGH	2026-09-05	Done 2026-09-05. Phase 2 ships as two releases: the Space Update (editors, the gate, other planets, inter-planet supply chains) and the War Update (1.0) (wars, factions, multiplayer, mobile). The old space/war boundary returns as a release schedule, not a design one: everything in Phase 2 may still assume everything else in Phase 2 exists, and the editor is built once, complete, in the Space Update. Recorded in 02_Scope & MVP Document v0.5, section Phase 2 Release Cadence. Was OI-350.
CI-055	docs	Eleven documents still describe a three-phase plan. 02_Scope & MVP Document collapsed to two phases on 2026-09-05 and 14_Win Conditions & Faction Allegiance was renumbered with it, but Phase 3 still appears in 01_Master GDD, 09_UI-UX Design Document, 06_Vertical Layers & Core Power System, 01_Technical Architecture Document, 02_Server Architecture Document, 02_Risk Register, 03_Monetization Design Document, ROADMAP and 01_World Bible. Every one of them means the war layer, which is now Phase 2, so the renumber is mechanical, but it is not blind: some rows say deferred to Phase 3 to mean not in the next release, and under two phases that sentence changes meaning rather than just its number. Needs a read of each, not a find and replace.	read each, do not sed it	the nine documents listed	-	HIGH	2026-09-05	Done 2026-09-05, and it was not a renumber. Reading each document found the word phase running on two unrelated axes: release phases (two, 02_Scope) and development phases (the build order, ROADMAP, whose M<phase>.<n> IDs are an interface and must never be renumbered), plus a third ad-hoc axis in 06_Vertical Layers. A find and replace would have corrupted ROADMAP and the milestone IDs the vault and its tooling reference. Fixed by adding 04_Data Dictionary §0a Phase Vocabulary and deprecating bare Phase N; renaming ROADMAP's headings to Development Phase N with a warning box; rewriting 06_Vertical Layers' Phase Scope as a Build Order that states the vertical system is release Phase 1 content; and repointing the genuine release references in 01_Master GDD, 09_UI-UX Design Document, 01_World Bible, 03_Monetization Design Document, 02_Risk Register, 01_Technical Architecture Document and 02_Server Architecture Document at the Space Update or the War Update. Raised OI-355. Was OI-354.
CI-056	docs	The Risk Register and the Scope document disagree about whether multiplayer is in release Phase 1. 02_Risk Register carried a decision row reading Full online multiplayer in Phase 1. Documented in 01_Technical Architecture Document, dated 2026-02-22. 02_Scope & MVP Document lists multiplayer or networking under What Is Explicitly NOT In Phase 1 and places it in the Phase 2 War Update. 01_Technical Architecture Document §Open Questions assumed multiplayer session discovery in Phase 1, which only makes sense under the Risk Register's reading. One of these is wrong and the difference is roughly a development phase of work (ROADMAP development phase 4). Surfaced by the OI-354 phase pass, which flagged both documents in place rather than picking a side. Note the reading is genuinely ambiguous at source: Phase 1 in the 2026-02-22 row may have meant a development phase, not a release phase, which is exactly the collision §0a now forbids.	El decides; do not infer it from either document	02_Risk Register decision log; 02_Scope & MVP Document What Is Explicitly NOT In Phase 1; 01_Technical Architecture Document section 400	-	HIGH	2026-09-05	Done 2026-09-05, resolved in the Risk Register's favour. Co-op is in release Phase 1; war multiplayer is in the Phase 2 War Update. The deciding evidence was ROADMAP, which has always sequenced development phase 4, Multiplayer, before development phase 5, Early Access Launch, so the February decision row was the standing plan and 02_Scope was the outlier. The root cause was the same overloading that caused OI-354: multiplayer is an umbrella covering two systems with two release dates. Fixed by adding 04_Data Dictionary §0b Multiplayer Vocabulary (co-op versus war multiplayer) and deprecating the bare word where the form matters. 02_Scope v0.6 adds co-op to Phase 1 Platform & Tech and to the Definition of Done, and rerates the scope risk from Medium to High, because co-op is now release-blocking rather than deferred. 02_Risk Register rows T2, T4 and D3 renamed to co-op, with D3 given a Phase 1 answer since its stated mitigation, role specialisation across Planet/Space/War, does not exist in Phase 1. 36_Progression Arc Design v0.2 adds §7a Beats in Co-op: a beat is world state and fires once for the world, everyone present takes the XP, a guest carries nothing in and the XP out. No new rule was needed. Was OI-355.
CI-057	pollution	Decide whether scavengers appear in PvE/peaceful sessions.	-	26_Pollution, Waste & Organic Systems Design -> Hostiles - Scavengers, Not Biters	-	NORMAL	2026-08-25	Done 2026-09-05. Yes, they appear in peaceful worlds, because they are drawn by the player's own choices rather than by a difficulty setting. E-waste dumped in the open and a high local power signature bring them; a clean, modest factory genuinely does not. Peace is earned, not toggled. Recorded in 37_Planet Threats & Defence Design section 4 and 6. Was OI-091.
CI-058	factions	Decide whether aliens are a negotiable fifth faction.	-	14_Win Conditions & Faction Allegiance -> Open Questions	-	DEFERRED	2026-03-31	Done 2026-09-05. Not a fifth faction, and not in Phase 1. The natives are a threat the player cannot talk to. What changed is that they are now machines, so who left these here becomes a real question, deliberately adjacent to 28_Restoration (derelicts with an invented history) and to The Remnants (deep-space ancient tech). Whether that becomes negotiable is a War Update question, and Early Access carries the mystery for free. Recorded in 37_Planet Threats & Defence Design section 3.1. Was OI-197.
CI-059	threats	Whether anything lives underground. 06_Vertical Layers gives each level its own atmosphere and 26_Pollution section 3a establishes that air pollution does not travel between levels. 37_Planet Threats & Defence Design section 7 uses that to produce there is no quiet level: burying dirty industry hides it from the natives at the cost of severe corrosion. That table assumes natives are a surface population. If anything nests underground the whole trade changes, and so does the value of the underground levels. Not a defect, a genuinely open design question that section 7 depends on.	answer before section 7 is built	37_Planet Threats & Defence Design section 7; 06_Vertical Layers & Core Power System	-	HIGH	2026-09-05	Done 2026-09-05. Nothing lives underground, and it is still not safe. The machines travel: natives dig, triggered by ground contamination leaching downward (26_Pollution section 3a.2a, downward only and slowly), and ferals follow the player's own shafts, above all the Power Conduit, drawn by power signature. Three consequences. Burying dirty industry hides the air but not the ground, which splits smelting and dumping into two separate decisions where they used to be one. The bill arrives late, because leaching is slow, landing after the player has built enough down there to care. And every shaft is a door, so shaft count becomes a defensive decision as well as a logistical one: fewer fatter shafts are cheaper to defend and a single point of failure, and a level with no conduit is unreachable and also unpowered. Section 7 rewritten as Underground Is Late, Not Quiet. Was OI-359.
CI-060	threats	Ammunition, weapons and turret behaviour need their own deep dive. Deliberately excluded from 37_Planet Threats & Defence Design, which says what is coming and why but not what you shoot it with. That conversation owns: whether turrets tier up or the ammunition does (the second is more in this game's grain and would change the current turret_basic / turret_laser pair in 04_Factory), what the ammunition production chain is, what the defence buildings actually are, how the Unit Spawner's deployable units work for clearing nests, and how walls and gates behave. Ammunition is one of Phase 1's three economy sinks (35_Economy Design), so this is not a polish item.	next design session after threats	37_Planet Threats & Defence Design section 11; 04_Factory Defence Buildings; 35_Economy Design	-	HIGH	2026-09-05	Done 2026-09-06. 38_Weapons & Ammunition Design v0.1 written. Defence is a production line, not a purchase, measured in items per minute. Two families: kinetic consumes items, energy consumes power, and energy defence raises the power signature that draws the ferals, deepening with Machine Mk since draw rises faster than output. Both the building and the ammunition tier, by different mechanisms: buildings by the four universal upgrade kits applied in place (Canonical section 4), ammunition by research recipes that require a Mk-N turret to chamber. So a defence line is never rebuilt to be improved, and it improves while unattended. One ammunition ladder, shared with ship weapons, drawn from goods the factory already makes, which makes the Phase 1 line the War Update line and closes half of OI-343. Magazine is a buffer so a dry line is a diagnosable logistics failure; turrets take modules and all four types have real meaning; ferals fly so walls answer natives, EMP answers ferals, turrets answer both; bots are built and spent, a batch sink against defence's rate sink. Was OI-360.
CI-061	research	Branch 6 unlocks two buildings that should be Mk levels. 08_Research Tree Design Branch 6 has Turret Upgrade II unlocks Heavy Turret (2x2, distinct building) and Advanced Fortification unlocks Reinforced Wall, and its audit note defends them: "Reinforced Wall and Heavy Turret are distinct buildings rather than Mk levels, so they are legitimate research unlocks." That is no longer true. El decided 2026-09-06 that all buildings tier, which under Canonical section 4 means by upgrade kit, not by research. Per 38_Weapons & Ammunition Design section 9 those two nodes should unlock higher ammunition recipes instead, requiring a turret at Mk-N to chamber; the audit note must be withdrawn with them. Turret Upgrade I (flat range and fire rate) and the Combat Doctrine infinite node survive unchanged. A branch amendment under the document's own section 0.3 protocol, not a rebuild.	amend the branch, do not rebuild	08_Research Tree Design Branch 6; 38_Weapons & Ammunition Design section 9	-	HIGH	2026-09-06	Done 2026-09-06. Both become Mk levels. 08_Research Tree Design v0.7, Branch 6 amendment section 0.5 under the document's own section 0.3 protocol. The 2026-08-30 audit note defending Reinforced Wall and Heavy Turret as distinct buildings is struck through and withdrawn. Advanced Fortification now unlocks gate automation only; Turret Upgrade II unlocks Hull Plating ammunition, requiring a Turret at Machine Mk2 to chamber. Turret Upgrade I and the Combat Doctrine infinite node unaffected, no other branch touched. Recorded the general rule this exposed: under Canonical section 4, no building is ever a research unlock at a higher tier — research unlocks recipes, a recipe may require a Mk-N building, and the Mk comes from a kit. Any node anywhere in the tree that unlocks a bigger version of an existing building is stale by construction. Was OI-361.
CI-062	threats	The rungs of the ammunition ladder above Hull Plating. Branch 6 starts the ladder with existing manufactured goods, Metal Plates then Hull Plating, and 38_Weapons & Ammunition Design keeps that on purpose: ammunition is things the factory already makes, so defence competes with expansion for the same items. Open whether the upper rungs stay as existing goods or need dedicated items, which is a judgement about whether a plate is ammunition still reads sensibly at the top of the tree. Also open: whether a higher rung raises damage per shot or rate of fire, since damage keeps the line's throughput requirement stable while rate raises consumption too.	damage over rate, probably	38_Weapons & Ammunition Design section 4 and Open Questions	-	NORMAL	2026-09-06	Done 2026-09-06. A rung buys rate of fire, range and spread. Never damage. Damage per shot is constant across the whole ladder. Four consequences, recorded in 38_Weapons & Ammunition Design section 4.2: nothing goes obsolete, since a Mk1 turret on Metal Plates is slow and short-ranged but still kills what it hits; every rung raises consumption, so researching one without scaling the line leaves the player worse off than before, which is defence is measured in items per minute in its purest form; it couples the two tier tracks, because a faster round drains the magazine and demands a higher Base Mk, cascading one research node into a production demand and a construction demand, neither of which is a build-menu click; and range and spread are geometry, so a rung changes where turrets belong rather than how many. Spread is the anti-swarm answer and therefore points at the ferals. Left open: whether spread needs a cap, and whether Turret Upgrade I's flat bonus on the same two axes should shrink so the ammunition stays the dominant source. Was OI-362.
CI-063	docs	Nine documents are linked but were never written. 23 wiki links across the vault point at documents that do not exist: 08_Economy & Progression Design (7 references), 06_Combat & War System Design (6), 07_Space Exploration Design (3), 03_Cross-Platform Strategy (2), 05_Spaceship Design & Building System, and 07_Reclaim in the Outpost folder, plus three unprefixed variants of the same names. None appears in the AGENTS.md Design Document Index, so they were planned and dropped rather than lost. Economy and Combat are the two that matter - seven and six documents each expect them to exist. Decide per document: write it, or remove the links. Do not create stubs, which would make the vault look complete while saying nothing.	write or delink; never stub	AGENTS.md -> Design Document Index	-	HIGH	2026-08-29	Done 2026-09-06. Written as 39_Space Layer Design v0.1, and Unwritten/07_Space Exploration Design converted to a redirect rather than deleted, since documents still link to that name. The stub's own excuse (space is Phase 2, so unwritten is expected rather than overdue) died on 2026-09-05 when the phase boundary moved and space became the second half of Early Access. Was OI-313.
CI-064	space	Genesis, and what it can and cannot do across planets. 39_Space Layer Design section 4 lets the player settle however they like — mines, one main factory, several mains, or a factory on every planet — and every one of those patterns depends on the builder. Open: whether Genesis travels, whether it can be in two places, whether a second one can exist, and whether a remote site can be built or maintained without it. El asked 2026-09-06 to take this as its own conversation rather than settle it inside the space layer, and section 4 is deliberately written to survive either answer. This is the largest undecided thing in the vault, because the answer changes what the space layer actually plays like.	El's own conversation; do not decide it inside another document	39_Space Layer Design section 4 and section 8; Research Branch 7 Genesis Abilities	-	HIGH	2026-09-06	Done 2026-09-07. 40_Genesis Design v0.1 written. Two rules were wearing one coat: indestructible is load-bearing (Canonical section 19, anti-softlock) and kept; always exactly 1, cannot be replicated is dropped, because it was what broke co-op and space. One Genesis per player. Bots are the builders; Genesis is the first one and the one that cannot die. Genesis founds a site, bots continue — a planet with no Construction Bot Bay needs Genesis to place the first one, after which bots build there unattended, which is exactly what 39_Space Layer Design section 4's settlement freedom needed. Founding is a trip that costs the home planet its ability to start anything new. Was OI-364.
CI-065	genesis	Decide whether multiple players can direct Genesis simultaneously or only through a queue.	-	-	-	NORMAL	2026-02-22	Done 2026-09-07. Neither: each player has their own Genesis. The question assumed one shared Genesis and asked whether players direct it simultaneously or through a queue. Under that rule, three of four co-op players cannot build, which El identified directly ("in co-op there is nothing for the other person to do"). Shared: the factory, the power grid, Factory Research and its slot pool, and construction materials. Personal: your Genesis and your build queue. Materials being shared is kept as a feature — two builders drain the plate line faster, which is a conversation rather than a rule. 40_Genesis Design section 3. Was OI-217.
CI-066	genesis	Design Genesis behaviour beyond the current task queue.	-	-	-	DEFERRED	2026-08-25	Done 2026-09-07, and the answer is that there is nothing to design. 04_Factory asserted that Bot Programming II makes Genesis graduate to a supervisory role (oversees bot fleet) without saying what that meant. Nothing supervises. The build queue does. Bots take work from the queue and so does Genesis; Genesis is simply the worker that was there first and cannot be destroyed. No supervision layer, no bot management screen, and no reason for one — inventing a management system for the thing that was meant to end management would undo the point. 40_Genesis Design section 4.1. Was OI-135.
CI-067	genesis	Whether founding a site really requires Genesis, or a freighter can deliver a self-unpacking Construction Bot Bay. 40_Genesis Design section 5 is the one call in that document Claude made rather than El, and it is flagged as such in place. Genesis founds, bots continue: a planet with no bay needs Genesis to land once, which gives interplanetary travel a decision (while your Genesis is away, home cannot start anything new where no bots reach) and keeps Genesis permanently relevant without making it the bottleneck. The alternative — a starter package the freighter drops — is simpler and costs the space layer that moment, since Genesis would then never need to travel. Matters most for solo play, where there is exactly one Genesis.	El confirms or flips; section 5 flags it	40_Genesis Design section 5; 39_Space Layer Design sections 4 and 8	-	HIGH	2026-09-07	Done 2026-09-07, confirmed and generalised by El: "Genesis has to be around for any build if there are no other bots." So the rule is not about founding specifically — Genesis builds wherever bots cannot, and founding a planet is simply the most obvious case. The freighter-drops-a-self-unpacking-bay alternative is dropped. El added the second case in the same message: if the bots run out of energy, Genesis can still do the build, which is now 40_Genesis Design section 4.0, bots are faster, Genesis never stops. That connects the power system straight to construction: when the grid sheds, the bot bay stops and the bots go quiet, so a blackout demotes the player back to one pair of hands, which is what makes First Blackout land as a beat rather than a notification. Genesis itself never stops — it draws from the grid when there is one and carries a reserve that always trickles back, so it slows rather than idles, per the Canonical section 19 anti-softlock rule. Was OI-370.
CI-068	data	The ship editor is unspecified, and the bill of materials cannot be computed without it. 01_Master GDD describes a "modular ship editor, player selects ship class first, then builds within that class's constraints" and 02_Scope & MVP Document says "hull + component attachment", which reads spatial rather than as a flat parts list. Nothing settles it. The difference matters: if attachment position is real then a design can be invalid as well as expensive, the editor has to say why, and a saved design carries positions rather than counts. If it is a parts list, a design is a dictionary of component to quantity and validity is just class constraints. 30_Ship Construction System depends on this: it computes a bill by summing placed components, and cannot be implemented against an undefined design. Also decides what a saved blueprint contains.	decide spatial or parts-list before anything else here	30_Ship Construction System section 3, section 10; Unwritten/05_Spaceship Design & Building System; 01_Master GDD -> Spaceship Design & Building	-	HIGH	2026-09-01	Done 2026-09-03 by 34_Ship & Mothership Editor Design, which names it in its Closes header, and confirmed in 30_Ship Construction System section 10 (OI-344 closed). Register housekeeping only, caught 2026-09-07: the row was marked closed in two documents but never moved out of the open register, where it had been sitting as a stale HIGH for four days. The editor specifies the escort layer model, mothership cells, class constraints, validity rules and the data shape of a saved design, which is what the bill of materials computation needed. Was OI-344.
CI-069	docs	The ship component list is depended on by two documents and owned by neither. 15_Ship & Army Points Design lists the ten components with APR values; 30_Ship Construction System sums those same components to compute a bill of materials; 04_Factory & Resource System Design lists a different five-component vocabulary that the ship list never references. This is the exact shape of the failure the vault already had once: the Unwritten/05 stub warns that "two documents on the same subject is how the vault ended up with two disagreeing fluid tables". Decide which document owns the list, have the others reference it, and never restate it. Ownership probably belongs with 15_Ship & Army Points Design, since APR is a per-component property and the list is already there, but that is a call rather than an obvious answer. Related to OI-343, which will add recipes to the same components and must put them in the owning document.	one owner, others reference	15_Ship & Army Points Design; 30_Ship Construction System section 10; 04_Factory & Resource System Design	OI-343	NORMAL	2026-09-01	Done 2026-09-03 by 34_Ship & Mothership Editor Design, which takes ownership of the component catalogue, and confirmed in 30_Ship Construction System section 10 (OI-345 closed). Register housekeeping only, caught 2026-09-07, same as OI-344: closed in the documents, never moved out of the register. One owner, others reference. Was OI-345.
CI-070	logistics	Design barrels. A barrel turns a fluid into an item so belts, trains and ordinary drones can carry it. Nothing about barrels exists in the vault. This is a lever on the entire fluid story: barrelling makes pipes optional for low volume rather than mandatory for all fluid. Decide barrel capacity against pipe throughput deliberately, or barrels either trivialise pipes or are never worth using.	-	06_Design Session Decision Log -> Transport & Logistics	-	HIGH	2026-08-27	Done 2026-09-08. 41_Fluid Packaging & Carriage Design v0.1 written. A pipe is a pool, a barrel is a parcel. Key correction from El: molten metal is still a liquid, so the constraint was never is it liquid at ambient but can the container hold it at temperature. That makes the container ladder mirror the pipe ladder — Barrel, Insulated Barrel, Cryogenic Barrel — because it is the same physics. The trivialise-or-useless trap the item warned about is avoided two ways: the insulated tier is roughly the same research and material commitment as Insulated Pipe, so it is a different shape of infrastructure rather than a bypass; and the machines at each end are the throttle, not the belt, so a pipe wins at volume automatically and barrel capacity becomes a free parameter tunable for feel rather than a balance lever. Was OI-272.
CI-071	logistics	Add barrelling and debarrelling recipes, plus an empty-barrel item. A filled barrel is not a liquid in a different form - it is a steel container, an ordinary SOLID item that happens to hold liquid, so it rides belts for the same reason every other solid does. Barrelling needs no special machinery: coolant (LIQUID) + empty_barrel (SOLID) -> coolant_barrel (SOLID) and the reverse, with the relationship living in the recipe graph like every other production relationship. Still needs a decision on barrel capacity against pipe throughput (OI-272).	ordinary recipes, no new concepts	06_Design Session Decision Log -> Resource Data Model	-	HIGH	2026-08-28	Done 2026-09-08 in 41_Fluid Packaging & Carriage Design section 7.1, with the recipe shape unchanged from what this row already specified: coolant (LIQUID) + empty_barrel (SOLID) -> coolant_barrel (SOLID) and the reverse, no special machinery, the relationship living in the recipe graph. Barrels are a loop, not a line (El, 2026-09-08): empties return, and Canonical section 12b applies — closed loops are always lossy — so a barrel line carries an ongoing steel cost, which is the honest price of packaging over piping. Also settled: flammable fluids in barrels are a fire risk, so a fire destroys steel and product together. Capacity deferred to the data pass as a free parameter. Was OI-278.
CI-072	logistics	Specify the drone-making building and the fluid-carrier variant. The building needs a name, footprint, recipe and category. The fluid-carrier drone needs deciding against barrels: if an ordinary drone can carry a barrel, what does a dedicated fluid drone add - capacity, or the ability to skip barrelling entirely?	-	06_Design Session Decision Log -> Drones; 04_Factory -> Drones	OI-272	NORMAL	2026-08-28	Partially done and partially dropped, 2026-09-08. The fluid-carrier drone variant is removed (El: "Remove fluid drone. But train keep."). 19_Transport & Logistics section 3.6 says packaging is the only way a solid carrier moves a fluid, so a fluid drone would break that rule for no gain: an ordinary drone carries barrels. What replaces it is the tank wagon (41_Fluid Packaging & Carriage Design section 4), making rail the only carrier that moves fluid as fluid without a pipe — which does not compete with barrels, since rail cannot reach another planet. The drone-making building still needs a name, footprint, recipe and category, and that half is re-raised as OI-375 rather than closed here. Was OI-297.
OI-089	pollution	undecided	Decide air pollution dispersal model: radius falloff or coarse grid diffusion.	diffusion is more interesting and more expensive	26_Pollution, Waste & Organic Systems Design -> Three Pollutants, Three Media	-	HIGH	2026-08-25
OI-377	logistics	undecided	Whether a cylinder can hold a liquefied gas, or whether that is the Cryogenic Barrel's job. 41_Fluid Packaging & Carriage Design establishes liquids go in barrels, gases go in cylinders, with pressure rating on the cylinder as temperature rating is on the barrel. The two overlap on liquefied gases (Methane, Ammonia from Gas and Ice Giants per 05_World Generation & Planet Types), and only one should own that case or the same substance gets two containers. Matters in the Space Update rather than at launch, but deciding it late means retrofitting recipes.	one owner, not both	41_Fluid Packaging & Carriage Design sections 2.1 and 3; 05_World Generation & Planet Types	-	NORMAL	2026-09-08
OI-372	ui	task	Make Genesis the voice of the notification system. El decided 2026-09-07 that Genesis talks and makes sounds, confirming the dialogue 10_Tutorial & Onboarding Design already gives it. 40_Genesis Design section 7 proposes the consequence: Genesis should deliver 18_Notifications, because three of the five Phase 1 beats depend on the game telling the player something at the right moment and First Blackout is dead unless the shed tier and the deficit are named out loud. A notification with a character attached lands where a toast in the corner does not. In co-op it stays personal — your Genesis talks to you, not to the room, so four players do not share one nagging voice. Constraints stated in section 7.1: not chatty, not a hint system, and sounds carry most of it. 18_Notifications was last touched 2026-03-31 and predates every beat, so this is part of the interface pass.	part of the interface and feel pass	40_Genesis Design section 7; 18_Notifications; 10_Tutorial & Onboarding Design	-	HIGH	2026-09-07
OI-381	ui	task	Wind has to be visible, and it is not yet. 26_Pollution section 3b.3 ships wind in the Launch release for everyone, which makes an invisible mechanic load-bearing on the player's first farm. This is a debt the decision created, not a nice-to-have. Needs: a persistent wind indicator readable without opening anything; a tutorial beat naming wind before the first farm is placed; and the wind direction drawn in the pollution overlay (section 3b.5) so the downwind tail reads as a consequence rather than a bug. Also carries the overlay's own specification - pollution, sinks and native zones shown together - and the ⚠️ cap on the world haze, which must stay well below obscuring the factory or it gets switched off permanently and takes the feedback with it. Feeds the interface and feel pass.	the decision is paid for here or not at all	26_Pollution sections 3b.3 and 3b.5; HUD ; UI-UX; the tutorial	-	HIGH	2026-09-11
OI-386	ships	undecided	Whether escorts use the layer model or the same freeform grid as the mothership. 34_Ship & Mothership Editor Design gives escorts concentric layers and the mothership a freeform grid, and argues the split explicitly: freeform on an escort "would be decoration", because only combat performance judges an escort and evaluating geometry for combat needs a damage simulation that 31_War & Combat System Design rules out. That argument was correct on the information available and is now in question, for two reasons. First, a 2026-09-13 session produced an evaluator that is not combat: power and fuel routed through the hull as wires and pipes, so a layout is valid or invalid whether or not anything ever shoots at it. Second, ⚠️ the document already refutes itself - its mothership section computes directional exposure per face from a freeform grid with no simulation, which is precisely the thing the escort section says cannot be done. One spatial model or two is the real question, and two models on one object would be the worse answer either way.	one model, and connectivity is what judges it	34_Ship & Mothership Editor Design sections Two Editors, One Reason, Escort Ships: The Layer Model and What the Mothership's Geometry Controls	-	HIGH	2026-09-13
OI-391	ships	undecided	How mothership hulls are unlocked, now that they are ordinary hulls. Before v0.5 the mothership used frames unlocked through Engineering Research, which is account-scoped, while ship classes are gated by Shipyard Mk, which is world-scoped. 34_Ship & Mothership Editor Design v0.5 folds motherships into one hull catalogue, which puts both gates on the same list and makes the mismatch visible. ⚠️ An account-scoped unlock granting world capability would rub against 36_Progression Arc Design section 7 - levels never gate world content - and against Canonical section 20, capacity crosses, stock does not. The original intent was probably that a player's mothership travels with them between worlds, which may still be right for the commissioned instance while the hull itself is gated in-world like everything else. Decide it deliberately rather than inheriting it.	one gate for all hulls unless there is a stated reason	34_Ship & Mothership Editor Design -> The Hull Catalogue and Mothership Hulls; 36_Progression Arc Design section 7	-	HIGH	2026-09-13
OI-392	ships	undecided	Conduit throughput: does one wire carry any draw, or does it have a capacity? 34_Ship & Mothership Editor Design v0.5 makes conduits occupy cells, which is what gives the envelope meaning - a fully packed ship cannot be wired. Unspecified is whether a single run feeds anything on it or whether a heavy draw needs parallel runs. ⚠️ It decides how much the rule bites on a large hull: with unlimited throughput a mothership needs one thin spine and the tension mostly evaporates at the size where it should be greatest; with capacity, a big ship spends real volume on its nervous system. Same question for fuel lines.	capacity, or the rule stops mattering exactly where it should matter most	34_Ship & Mothership Editor Design -> Wires and Pipes	-	NORMAL	2026-09-13