fix(@projects/@magic-civilization): 🐛 update regression objective with latest status

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

.project/objectives/p2-10-regression-ci-gate.md

@@ -5,7 +5,7 @@ priority: p2
 status: partial
 scope: game1
 owner: testwright
-updated_at: 2026-05-07
+updated_at: 2026-05-14
 evidence:
   - .forgejo/workflows/ci.yml
   - .forgejo/RUNNER_SETUP.md
@@ -66,9 +66,26 @@ CI Stage 3 remains `continue-on-error: true` until file-splitting is scheduled.
 - Benchmark regression tracking (future).
 - Signed commits / tag verification (future; not on the critical path).
 
-## Remaining work (2026-05-04)
+## Remaining work (2026-05-14)
 
-Three un-met acceptance items remain. Status stays `partial` until all three flip.
+Four un-met acceptance items remain. Status stays `partial` until all four flip.
+
+**Workflow-composition status (2026-05-14):** `.forgejo/workflows/ci.yml` is
+complete — every child stage (Rust workspace, gpu-feature best-effort, gdlint,
+validate-game-data, objectives-report --check, build-gdext, godot --import,
+headless GUT, autoplay smoke, artifact upload) is present and triggers on
+`push: branches: [main]` + `workflow_dispatch`. The remaining blockers are
+upstream of yaml composition: they live in Rust crate sources, GDScript
+authoring, and a workflow-policy decision on `max-file-lines`. None of them
+are solvable by editing the workflow yaml.
+
+**Children fold-in:**
+
+- p2-10a (gdlint un-gate) — `done`
+- p2-10b (GUT un-gate) — `done`
+- p2-10c..p2-10j (regression-batch fixes spawned during un-gating) — all `done`
+- p2-10k (51→11 gdlint cleanup) — `partial`; blocks Stage 3 hard-green
+- p2-10l (15→3 GUT regression triage) — `partial`; blocks Stage 5 hard-green
 
 **Ground-truth correction (2026-05-04)**: The previous snapshot (2026-04-23)
 described 7 gdlint violations and 39 GUT failures. Verified on apricot
@@ -120,6 +137,33 @@ fail — the pipeline is red. Tracking children p2-10k (gdlint) and p2-10l
   systemctl --user enable --now forge-watch
   ```
 
+### 4. Stage 1 hard-fails — blocked on p1-42
+
+- **Bullet**: ✓ hard — `cargo test --workspace --locked` (Stage 1) — *workflow
+  stage is wired and required; underlying compile is currently red on main.*
+- **Tracking parent**: `p1-42-ai-full-building-catalog.md` (status `partial`)
+- **Verified 2026-05-14** locally on mac (`cargo check -p mc-player-api`):
+  ```
+  error[E0063]: missing field `building_priors` in initializer of
+                `TacticalPlayerState`
+      --> crates/mc-player-api/src/projection.rs:1138:5
+  ```
+- **Root cause**: p1-42 added a `building_priors` field to `TacticalPlayerState`
+  in `mc-player-api/src/state.rs` (the AI-side replacement for the hardcoded
+  8-id ladder) but the `TacticalPlayerState { ... }` constructor in
+  `mc-player-api/src/projection.rs::project_player` at line 1138 was not
+  updated to populate it. The struct literal therefore fails E0063.
+- **Resolution path**: a single one-line addition inside `project_player`
+  populating `building_priors` from the player projection — owned by p1-42
+  (AI / mc-player-api), out of scope for this objective and explicitly
+  out of scope for the simulator-infra rotation that filed this audit.
+- **Implication for the CI gate**: until p1-42's `building_priors` plumbing
+  is finished, every push to main will fail at Stage 1 (`cargo test
+  --workspace --locked`) before any of the GDScript stages even run. The
+  workflow composition is correct; the underlying source is red. Do NOT
+  scope-narrow the workflow to compiling crates only — that would mask a
+  real production breakage on main and violate Rail-1.
+
 ### Runtime budget (information-only)
 
 - **Bullet**: 🟡 Runtime budget ≤15 min median — partial data.