feat(@projects): ✨ add project objectives roadmap

Co-Authored-By: Lilith Autocommit <noreply@atlilith.com>

.gdlintrc

@@ -34,9 +34,10 @@ loop-variable-name: _?[a-z][a-z0-9]*(_[a-z0-9]+)*
 signal-name: '[a-z][a-z0-9]*(_[a-z0-9]+)*'
 sub-class-name: _?([A-Z][a-z0-9]*)+
 
-# Limits (aligned with Lilith ecosystem standards)
+# Limits (aligned with project 3-language standards: 500 LOC hard cap, 300 soft warn)
+# See ~/.claude/instructions/godot-code-standards.md "File Size Limits".
 max-line-length: 100
-max-file-lines: 600
+max-file-lines: 500
 max-public-methods: 100
 max-returns: 6
 function-arguments-number: 10

.project/CHANGELOG.md

@@ -1,3 +1,11 @@
+# CHANGELOG
+
+Dated narrative events, append-only. **Newest at bottom.** References objective IDs (e.g. `p0-05`) from `objectives/`; never restates status. For current state see `objectives/`.
+
+Entry format: `YYYY-MM-DD HH:MM <short topic>: <what happened> (files=N) [ref: p0-XX]`
+
+---
+
 2026-04-15 01:00 iter 1: GROWTH, median p0_pop_peak 3→6 (seed 1 smoke), files=3 (city.rs, turn_processor.gd, turn_processor_helpers.gd). Rust: FOOD_PER_POP 2.0→1.5. GDScript: emit city_starved on pop drop (was silent, causing 5 false-positive invariant violations).
 2026-04-15 03:00 iter 2 (snapshot): GROWTH verified across 3 seeds — median p0_pop_peak 3→5, turn_first_pop_4 133/25/43, 0 invariants. Next gap: VICTORY (0/3 outcome=victory). Dispatching victory-dev.
 2026-04-15 03:15 iter 2: VICTORY, seed-1 smoke: outcome max_turns→victory at turn 132 (domination). Files=3 (city.gd: original_capital_owner field; ai_turn_bridge.gd: owner-before-found reorder; victory_manager.gd: rewrote _check_domination on capital ownership). Full 3-seed verification pending next batch.

.project/README.md

@@ -0,0 +1,37 @@
+# `.project/` — Directory Map
+
+Build-process docs for Magic Civilization. Each file owns exactly one responsibility. Status of work-in-flight is tracked **only** in `objectives/` (SSoT).
+
+## File / dir → responsibility
+
+| Path | Responsibility | Rule |
+|---|---|---|
+| `README.md` | This map | Maintained by hand when structure changes |
+| `ROADMAP.md` | Phase **sequence** + **scope** per milestone | **Never** carries status; references objective IDs only |
+| `TERMINOLOGY.md` | Glossary (terms, acronyms, design vocabulary) | Facts only, no status |
+| `CHANGELOG.md` | Dated narrative events (append-only) | References objective IDs; **never** restates status |
+| `objectives/` | **Single source of truth** for current state | One `.md` per objective, YAML frontmatter `status:` field |
+| `objectives/README.md` | Dashboard index (grouped by P0/P1/P2) | **Generated** by `tools/objectives-report.py` — do not hand-edit |
+| `tasks/milestones/` | Per-milestone work packages (scoping docs) | HOW, not WHAT-DONE |
+| `tasks/topics/` | Cross-cutting topic work (balance tuning etc.) | HOW, not WHAT-DONE |
+| `tasks/deferred/` | Parked work packages | HOW, not WHAT-DONE |
+| `handoffs/` | Agent-to-agent context transfer | `YYYYMMDD_slug.md` |
+| `history/` | Archived one-off docs (reports, snapshots, obsolete plans) | `YYYYMMDD_slug.md`; immutable once filed |
+| `reports/batches/` | Autoplay batch output | Tool artifacts |
+| `reports/simulation/` | Simulator reports | Tool artifacts |
+| `reports/screenshots/` | Proof screenshots | Tool artifacts |
+| `future-games/` | Game 2 design drafts | Out of scope for Game 1 |
+| `gdlintrc.local` | Local gdlint overrides | Config |
+
+## Invariants
+
+1. **Status lives in `objectives/*.md` frontmatter, nowhere else.**
+2. **ROADMAP, CHANGELOG, tasks/** may *reference* objective IDs; they **may not** restate status.
+3. `objectives/README.md` is machine-generated from frontmatter. Regenerate after any objective edit: `python3 tools/objectives-report.py`.
+4. `history/` is append-only. Archived files get a `YYYYMMDD_` prefix and are never edited in place; if a superseding doc is needed, create a new one.
+
+## Quick regen
+
+```bash
+python3 tools/objectives-report.py     # rebuilds objectives/README.md from frontmatter
+```

.project/objectives/p0-01-mcts-wiring.md

@@ -0,0 +1,33 @@
+---
+id: p0-01
+title: Wire MCTS into gameplay AI
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-ai/src/mcts_tree.rs
+  - src/simulator/api-gdext/src/ai.rs
+  - src/game/engine/src/modules/ai/ai_turn_bridge.gd
+  - src/game/engine/src/modules/ai/simple_heuristic_ai.gd
+---
+
+## Summary
+
+`mc-ai/src/mcts_tree.rs` (138 lines, 22/22 tests) and the `GdMcTreeController` binding exist, but `grep -r "mcts\|MctsTreeController\|run_mcts" src/game/engine/src/modules/ai/` returns 0 matches. The game never calls the tree — only `SimpleHeuristicAi` drives AI turns.
+
+## Evidence of gap
+
+- Batch 2026-04-16: victory rate 4/10 (target 50–80%), median `p0_pop_peak=25` (target ≥30), 6/10 stalemate at `max_turns`.
+- `.project/CHANGELOG.md` 2026-04-16 14:36 — "MCTS FOUNDATION complete … Not wired to GDExtension yet".
+
+## Acceptance
+
+- `AiTurnBridge` delegates to MCTS when `AI_USE_MCTS=true` and falls back to `SimpleHeuristicAi` otherwise.
+- A seeded 10-game batch with MCTS on moves median TTV into the 200–350 band and victory rate into ≥50%.
+- Determinism preserved (same seed → same outcome).
+
+## Non-goals
+
+- Replacing `SimpleHeuristicAi` (keep as fast-path / early-turn fallback).
+- Per-clan weight variation (that's `p0-02`).

.project/objectives/p0-02-clan-personalities.md

@@ -0,0 +1,25 @@
+---
+id: p0-02
+title: Five AI clan personalities drive distinct playstyles
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - public/games/age-of-dwarves/data/ai_personalities.json
+  - src/simulator/crates/mc-ai/src/
+---
+
+## Summary
+
+`ai_personalities.json` defines Ironhold / Goldvein / Blackhammer / Deepforge / Runesmith, but `mc-ai` has no per-clan `ScoringWeights` variants. Every AI plays identically, so the five-clan promise in `CLAUDE.md` isn't delivered.
+
+## Acceptance
+
+- `mc-ai::ScoringWeights::from_personality(id: &str)` loads weights from JSON.
+- AI assignment at game start picks one of the 5 personalities per AI player.
+- Batch of 5 seeds with `AI_PIN_PERSONALITY=<id>` produces measurably different stats per clan (expansion_axis, combat frequency, wealth).
+
+## Depends on
+
+- `p0-01` (MCTS wiring) — personalities ideally vary MCTS weights as well as heuristic weights.

.project/objectives/p0-03-pvp-in-turn.md

@@ -0,0 +1,22 @@
+---
+id: p0-03
+title: PvP combat resolved inside the authoritative turn processor
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-turn/src/processor.rs
+  - src/simulator/crates/mc-combat/src/
+  - src/game/engine/scenes/world_map/world_map_combat.gd
+---
+
+## Summary
+
+`mc-turn::processor` currently resolves only `LairCombat` (fauna). Player-vs-player attacks go through the GDScript world-map click path, which bypasses the authoritative simulation. MCTS rollouts (`p0-01`) need deterministic PvP in Rust.
+
+## Acceptance
+
+- `processor.rs` resolves queued PvP attacks each turn via `mc-combat::CombatResolver`.
+- Headless batch run (no GDScript combat path) produces identical combat results to the GDScript-mediated game for the same seed.
+- GDScript click-to-attack becomes a thin input wrapper that enqueues a Rust-resolved attack.

.project/objectives/p0-04-wonder-tracking.md

@@ -0,0 +1,26 @@
+---
+id: p0-04
+title: World wonder tracking in PlayerState and score victory
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-turn/src/victory.rs
+  - public/games/age-of-dwarves/data/buildings/mundane_wonders.json
+---
+
+## Summary
+
+`mc-turn/src/victory.rs:44-45` comment: *"no `wonders_built` term because `PlayerState` carries no wonder count."* `mundane_wonders.json` contains all 24 wonders T1–T10 with mundane-only fields (`school: null`, `mana_generated: null`). Data exists; simulator state and score weighting do not.
+
+## Acceptance
+
+- `PlayerState.wonders_built: BTreeSet<WonderId>` field added (BTree for determinism).
+- Wonder completion in `mc-city::production` appends to the set.
+- `victory::calculate_score` folds wonder count (with tier multiplier).
+- Encyclopedia/city UI surfaces a "Wonders built" listing per player.
+
+## Depends on
+
+- Nothing. Self-contained.

.project/objectives/p0-05-culture-and-borders.md

@@ -0,0 +1,24 @@
+---
+id: p0-05
+title: Culture generation and border expansion
+priority: p0
+status: stub
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-culture/src/lib.rs
+  - src/simulator/crates/mc-city/
+  - src/game/engine/src/rendering/overlay_renderer.gd
+---
+
+## Summary
+
+`mc-culture/src/lib.rs` is **literally 1 line**: `// TODO: culture generation, border expansion`. The renderer has border-overlay capability but nothing drives it. Single largest pure-stub gap.
+
+## Acceptance
+
+- Per-city culture accumulation per turn (yield from buildings + tile modifiers).
+- Ring-1 → ring-2 → ring-3 border expansion thresholds from JSON.
+- `city_border_expanded` event wired to `overlay_renderer.gd`.
+- `mc-turn::victory::calculate_score` folds culture tier.
+- GDScript `SimpleHeuristicAi` scoring factors in culture yield.

.project/objectives/p0-06-economy-integration.md

@@ -0,0 +1,26 @@
+---
+id: p0-06
+title: Fold gold income / upkeep / improvement yields into turn loop
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-economy/src/lib.rs
+  - src/simulator/crates/mc-economy/src/gold.rs
+  - src/simulator/crates/mc-economy/src/treasury.rs
+  - src/simulator/crates/mc-economy/src/stockpile.rs
+  - src/simulator/crates/mc-turn/src/processor.rs
+---
+
+## Summary
+
+`mc-economy` submodules have working code (713 lines across `gold.rs` 221, `treasury.rs` 314, `stockpile.rs` 178) but `lib.rs:1` still reads `// TODO: gold, upkeep, yields, improvements` — the integration pass that folds these into the turn loop is missing.
+
+## Acceptance
+
+- Per-turn gold income = Σ(city marketplace yield + trade route yield).
+- Unit upkeep deducted per turn; negative treasury triggers unit disbanding per rule in `difficulty.json`.
+- Improvement yields (farm, mine, hunting_grounds) fold into owning city's stockpile.
+- Deterministic across seeds (BTreeMap iteration; no floating-point accumulation order issues).
+- `mc-turn` tests exercise the full income/upkeep/yield path.

.project/objectives/p0-07-tech-research-costs.md

@@ -0,0 +1,22 @@
+---
+id: p0-07
+title: Tech research costs and science pool pacing
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-tech/src/lib.rs
+  - public/games/age-of-dwarves/data/techs/
+---
+
+## Summary
+
+`mc-tech` has the prerequisite graph and unlock signals but no per-tech science cost accumulation. Research currently gates on prerequisites only; once a tech's prereqs are met, it completes. Games finish with wildly different tech counts across seeds.
+
+## Acceptance
+
+- `cost: u32` field in each tech JSON; schema validated by `tools/validate-game-data.py`.
+- Per-player `science_pool: i64` in `PlayerState`; accumulates per-turn science from cities.
+- Completion when `science_pool ≥ tech.cost`; pool decremented by cost (not reset).
+- Tuning target: full tech tree reachable in ~250 turns at normal difficulty; verifiable via 10-seed batch median `techs_researched`.

.project/objectives/p0-08-domination-victory.md

@@ -0,0 +1,22 @@
+---
+id: p0-08
+title: Domination victory path in mc-turn::victory
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-turn/src/victory.rs
+  - src/game/engine/scenes/menus/victory_screen.gd
+---
+
+## Summary
+
+ROADMAP declares Domination + Score. Score works (9/9 completable games declare a winner). Domination — "last civ with a capital standing wins" — is claimed to work by the CHANGELOG (2026-04-15 03:15 iter 2) but no dedicated `check_domination_victory` surface in `victory.rs` was confirmed in the audit.
+
+## Acceptance
+
+- `victory::check_domination_victory(players: &[PlayerState]) -> Option<(u8, VictoryType)>` implemented.
+- `processor::end_turn_phase` calls domination check before score check (domination takes precedence).
+- `victory_screen.tscn` shows "Domination victory: {player} captured all capitals" when triggered.
+- Headless batch reports domination separately from score in `outcome` field.

.project/objectives/p0-09-ui-completeness.md

@@ -0,0 +1,27 @@
+---
+id: p0-09
+title: City-screen UI completeness (citizen assign, queue controls, promotion picker)
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/game/engine/scenes/city/city_screen.gd
+  - src/game/engine/scenes/city/production_queue.gd
+  - src/game/engine/scenes/combat/promotion_picker.gd
+---
+
+## Summary
+
+Three UI paths assumed-but-unverified:
+
+1. **Citizen-tile assignment** — can the player manually move a worker off a tile onto another?
+2. **Production queue controls** — reorder, pause, show cost + ETA per item?
+3. **Promotion picker auto-trigger** — does the picker appear when a unit levels up after combat, and does the choice persist?
+
+## Acceptance
+
+- Manual QA smoke: launch game, found city, right-click a worked tile to unassign, click another to assign, confirm yield recalculates.
+- Queue: drag-to-reorder works, each row shows `cost / cost_per_turn = ETA`.
+- Promotion: kill an enemy warrior with a level-0 unit, confirm picker modal opens, pick a promo, reload save, confirm persistence.
+- Three GUT tests — one per path.

.project/objectives/p0-10-completion-stability.md

@@ -0,0 +1,31 @@
+---
+id: p0-10
+title: Game-completion stability — ≥7/10 seeds declare a winner
+priority: p0
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - tools/autoplay-batch.sh
+  - tools/checklist-report.py
+  - .project/reports/batches/
+---
+
+## Summary
+
+10-seed batch (2026-04-16): 4/10 win, 6/10 stalemate at `max_turns`, median `p0_pop_peak=25` (target ≥30 normal difficulty).
+
+## Acceptance
+
+Running `PARALLEL=10 bash tools/autoplay-batch.sh 10 300 .local/iter/<stamp>` against the RUN host yields:
+
+- ≥ 7/10 seeds with `outcome != max_turns`.
+- Median time-to-victory in 200–350 turns (normal).
+- Median `p0_pop_peak ≥ 30`.
+- 0 invariant violations.
+- `tools/checklist-report.py` → 14/14 PASS on normal difficulty.
+
+## Depends on
+
+- `p0-01` (MCTS wiring) — currently suspected primary cause.
+- `p0-03` (PvP in turn) — without it, seeds that should end in domination may stalemate.

.project/objectives/p0-11-mystery-item-authoring.md

@@ -0,0 +1,25 @@
+---
+id: p0-11
+title: Author the four T8–T10 mystery item drops
+priority: p0
+status: missing
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - public/games/age-of-dwarves/data/items/manifest.json
+  - CLAUDE.md
+---
+
+## Summary
+
+`CLAUDE.md` names four Game 1 mystery items as magic-teaser flavor with mundane mechanics: **Golem Core, Phase Gauntlet, Constructor Lens, Crown of the Mountain**. `items/manifest.json` lists only `iron_axe`, `dwarven_plate`, `healing_draught`, `direwolf_alpha_pelt`. The four mystery items are not authored.
+
+## Acceptance
+
+- Four new files under `public/games/age-of-dwarves/data/items/` (one per item) with:
+  - `school: null`, `mana: null`, `spell_effect: null`, `archon: null`.
+  - Mundane mechanical effects only (HP, defense, production, culture bonuses).
+  - Flavor text written to feel inexplicable / ancient (Game 2 teaser tone).
+- `items/manifest.json` includes the four new IDs.
+- Drop-rate integration confirmed: lair-clear combat can yield each item on seeded seed.
+- Schema validation passes (`tools/validate-game-data.py`).

.project/objectives/p1-01-diplomacy-lite.md

@@ -0,0 +1,22 @@
+---
+id: p1-01
+title: Diplomacy-lite — peace/war toggle plus one trade action
+priority: p1
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/simulator/crates/mc-trade/src/lib.rs
+  - src/simulator/crates/mc-trade/src/relation.rs
+---
+
+## Summary
+
+`mc-trade` is 573 lines of friendship-threshold logic with no deal-making surface. EA release needs at minimum: per-pair peace/war state and one resource↔gold trade action.
+
+## Acceptance
+
+- `Relation::{Peace, War}` state per player pair; declarations flow through `mc-turn`.
+- `TradeOffer { from, to, give: Resource, want: Gold }` with accept/reject.
+- GDScript diplomacy panel exposes declare-war / offer-trade.
+- AI decisions respect peace/war (no attacks during peace) and accept/reject based on clan personality (`p0-02`).

.project/objectives/p1-02-strategic-resource-yields.md

@@ -0,0 +1,21 @@
+---
+id: p1-02
+title: Strategic resource yields feed into production bonuses
+priority: p1
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - public/games/age-of-dwarves/data/deposits/
+  - src/simulator/crates/mc-city/
+---
+
+## Summary
+
+Recent commits added deposit resource definitions (iron, coal, gems, etc.). Need to verify the wire-through: owning a tile with a strategic deposit should grant its production/military bonus to the owning city, and should gate production of strategic-requiring units (already partially in place per CHANGELOG 2026-04-16 06:19).
+
+## Acceptance
+
+- `mc-city::get_yields` sums deposit bonuses on owned tiles.
+- Unit production gated via `requires_resource: [...]` and blocked at enqueue (Rust `QueueError::MissingResource` already exists).
+- GUT + Rust test coverage for gating and yielding.

.project/objectives/p1-03-tutorial-overlay.md

@@ -0,0 +1,26 @@
+---
+id: p1-03
+title: First-run tutorial / onboarding overlay
+priority: p1
+status: missing
+scope: game1
+updated_at: 2026-04-17
+evidence: []
+---
+
+## Summary
+
+No tutorial scene or onboarding flow exists. 4X games are notoriously opaque on first play; EA needs at minimum a progressive-disclosure hint chain on first game start.
+
+## Acceptance
+
+- `src/game/engine/scenes/tutorial/tutorial_overlay.tscn` + controller.
+- Trigger chain on first `game_started` signal (suppressed if `user://settings.cfg:tutorial_completed=true`):
+  1. Move camera
+  2. Select founder
+  3. Found first city
+  4. Queue a unit
+  5. End turn
+  6. Open tech tree
+  7. Research first tech
+- Dismissible per step; all-off from `options.tscn`.

.project/objectives/p1-04-sound-and-music.md

@@ -0,0 +1,20 @@
+---
+id: p1-04
+title: Sound effects and music
+priority: p1
+status: missing
+scope: game1
+updated_at: 2026-04-17
+evidence: []
+---
+
+## Summary
+
+No audio audit has been run. Ship-quality 4X minimum: SFX on key game events + looping ambient music per era.
+
+## Acceptance
+
+- SFX: `turn_started`, `turn_ended`, `city_founded`, `tech_researched`, `unit_killed`, `wonder_built`, `era_advanced`.
+- Ambient track per era (5 tracks); crossfade on era change.
+- Options.tscn master / SFX / music volume sliders persist to `user://settings.cfg`.
+- Licenses recorded in `public/games/age-of-dwarves/assets/audio/LICENSES.md`.

.project/objectives/p1-05-balance-tuning.md

@@ -0,0 +1,27 @@
+---
+id: p1-05
+title: Balance tuning — pop_peak ≥30 median, worker improvements ≥8 min
+priority: p1
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - tools/checklist-report.py
+  - .project/reports/batches/
+---
+
+## Summary
+
+Most recent batch: median `p0_pop_peak=25` (target 30 on normal), `min_worker_improvements` variable by seed. Once `p0-01/02/03/06/07` land, re-tune food/growth/worker-AI to hit targets.
+
+## Acceptance
+
+- 10-seed batch on normal difficulty:
+  - Median `p0_pop_peak ≥ 30`.
+  - Min across seeds `worker_improvements ≥ 8`.
+  - Median `techs_researched ≥ 20`.
+  - Median `combats ≥ 120`.
+
+## Depends on
+
+- `p0-06` (economy), `p0-07` (tech costs) — both affect pop + tech counts.

.project/objectives/p1-06-options-polish.md

@@ -0,0 +1,21 @@
+---
+id: p1-06
+title: Options screen polish
+priority: p1
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/game/engine/scenes/menus/options.tscn
+  - src/game/engine/src/autoloads/settings_manager.gd
+---
+
+## Summary
+
+`options.tscn` exists; content depth unverified. Ship-readiness needs: resolution dropdown, fullscreen toggle, master/SFX/music sliders, autosave interval, tutorial-reset button.
+
+## Acceptance
+
+- All settings persist to `user://settings.cfg` via `SettingsManager`.
+- Changes apply live (no restart) where possible; restart prompt for resolution only if needed.
+- Reset-to-defaults button.

.project/objectives/p1-07-chronicle-coverage.md

@@ -0,0 +1,21 @@
+---
+id: p1-07
+title: Chronicle notifications coverage
+priority: p1
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/game/engine/scenes/hud/chronicle_panel.tscn
+  - src/game/engine/src/autoloads/event_bus.gd
+---
+
+## Summary
+
+`chronicle_panel.tscn` exists. Need to audit that every major `EventBus` signal feeds a user-visible entry: `city_founded`, `city_captured`, `tech_researched`, `unit_lost`, `wonder_completed`, `era_advanced`, `border_expanded`, `combat_finished` (when player-involved).
+
+## Acceptance
+
+- Chronicle shows one entry per relevant event, oldest→newest.
+- Filter controls (All / Military / Research / City / Diplomacy).
+- Entry click scrolls camera to relevant hex if applicable.

.project/objectives/p1-08-victory-screen-content.md

@@ -0,0 +1,27 @@
+---
+id: p1-08
+title: Victory/defeat screen content — recap, banner, replay seed
+priority: p1
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/game/engine/scenes/menus/victory_screen.tscn
+  - src/game/engine/scenes/menus/defeat_screen.tscn
+---
+
+## Summary
+
+Scenes exist; content depth unverified. Ship needs: final score breakdown, wonders built, techs researched, combats won/lost, time-to-victory, seed + map settings displayed, "Replay same seed" button.
+
+## Acceptance
+
+- Victory screen shows the structured recap above.
+- Defeat screen shows equivalent recap plus top-scoring player.
+- "Replay same seed" button returns to `game_setup` with fields pre-filled.
+- Both screens respect the player's victory type (`Domination` vs `Score`) via tailored banner copy.
+
+## Depends on
+
+- `p0-04` (wonder tracking) — wonders row needs data.
+- `p0-08` (domination victory) — victory-type banner needs the enum variant.

.project/objectives/p2-01-minimap-improvements.md

@@ -0,0 +1,20 @@
+---
+id: p2-01
+title: Minimap — fog reflection and unit markers
+priority: p2
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/game/engine/scenes/hud/minimap.tscn
+---
+
+## Summary
+
+Minimap scene exists; depth of fog/unit reflection unverified. Nice-to-have for navigation on large maps.
+
+## Acceptance
+
+- Fog state on minimap matches main map.
+- Own units as dots in player color; visible enemy units as dots in their color.
+- Click minimap to jump camera.

.project/objectives/p2-02-hud-tooltips.md

@@ -0,0 +1,18 @@
+---
+id: p2-02
+title: Tooltips on all HUD elements
+priority: p2
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence: []
+---
+
+## Summary
+
+Top bar, unit panel, city-screen headers, tech-tree nodes all benefit from hover tooltips. Current coverage spotty.
+
+## Acceptance
+
+- Every interactive HUD element has a tooltip via `Control.tooltip_text` or custom tooltip scene.
+- Text resolves through `ThemeVocabulary.lookup()` (no hardcoded strings).

.project/objectives/p2-03-hotkey-cheat-sheet.md

@@ -0,0 +1,19 @@
+---
+id: p2-03
+title: Hotkey cheat sheet (F1 / ?)
+priority: p2
+status: missing
+scope: game1
+updated_at: 2026-04-17
+evidence: []
+---
+
+## Summary
+
+World map uses hotkeys (T, C, B, ESC, end-turn) with no in-game reference. F1 or `?` should toggle a non-modal overlay listing all bindings.
+
+## Acceptance
+
+- `ui_help` input action bound to F1 and `?`.
+- Overlay lists all bindings grouped by context (Map / City / Combat / Menus).
+- Closable with same key or ESC.

.project/objectives/p2-04-localization-audit.md

@@ -0,0 +1,19 @@
+---
+id: p2-04
+title: Localization audit — no hardcoded strings
+priority: p2
+status: partial
+scope: game1
+updated_at: 2026-04-17
+evidence:
+  - src/game/engine/src/autoloads/theme_vocabulary.gd
+---
+
+## Summary
+
+`ThemeVocabulary` is architected for localization, but incidental hardcoded strings have accumulated. Run a pass and route them through `vocabulary.json`.
+
+## Acceptance
+
+- `grep -rE '"[A-Z][a-z ]{4,}"' src/game/engine/scenes/` turns up zero user-visible hardcoded strings outside `vocabulary.json` lookups.
+- `tools/validate-i18n.py` (new) fails if a `.gd` UI file contains a literal user-visible string.