phase: 03_simulation_movement_and_interaction

**Final verification pass — phase 03 `simulation_movement_and_interaction` (all 5 task files already in `complete/`)**

- Verified all phase-03 deliverables exist and work: `sim/` kernel (`step`, `state_build`, `rng`, `movement`, `collision`, `interact`, `carry`, `replay`), replay fixtures + 3 committed goldens, `docs/SIMULATION.md`.
- `bash .agents/validate.sh` → **passed**: ruff `All checks passed!`, pyright `0 errors`, `undercooked-levels validate` green, full `pytest` **1033 passed** (coverage **99.71%** ≥ 90% floor), slow sim stage green, repo-hygiene clean.
- Perf smoke (4 players, 2 000 ticks): mean **0.049 ms/tick, 20 423 ticks/s** — ≥ 10 000 ticks/s criterion met (test lives in `test_perf_smoke.py`, which the gate's slow stage runs; the overview names it `test_replay.py`).
- Completion criteria: import `undercooked.sim.step.step` ✅ · `tests/sim` green incl. determinism + goldens ✅ · perf ✅ · constant-change-fails-by-design ✅ (committed-hash equality + mutation test naming first bad tick/field; goldens never rewritten by tests) · `docs/SIMULATION.md` ✅ (constants, 8-step order, purity/determinism, re-bless) · full suite + coverage ✅ · app launches: `/healthz` 200, `/` 200, `/api/levels` serves content ✅ · replay CLI `check walk_solo.json` exit 0 ✅ · no regression in phases 01/02 (their suites green) ✅.
- Documentation: README.md — fixed one stale inconsistency (Status opening listed only 01/02 complete with 03 pending; now lists 01–03 complete, next pending `04_kitchen_pipeline_and_scoring`); all other README claims (CLI flags, config table, shell text, replay CLI exit codes) verified against code. AGENTS.md — verified consistent with the project as it stands, no changes needed.
- No deviations. Next pending phase: **`04_kitchen_pipeline_and_scoring`**.

docs/TESTING.md

@@ -20,10 +20,10 @@ layer live in `tests/conftest.py`, and layer-specific fixtures live in that laye
 | Directory | Contains | I/O allowed |
 |---|---|---|
 | `tests/unit/` | pure logic with no I/O: config parsing, header values, CLI precedence, protocol/validator helpers, and the doc/content guards that keep README, `.env.example`, `PLAN.md` and the code agreeing | none (in-process `TestClient` requests are fine) |
-| `tests/sim/` | tests of the pure kernel that need no server: grid geometry, the entity model, the A5 import-purity guard, and the level validator (phase 02) — plus the headless determinism/replay suites with golden state hashes (`tests/sim/test_replay_*.py`) that arrive with the tick in phase 03 | none — the sim must stay callable without a server |
+| `tests/sim/` | tests of the pure kernel that need no server: grid geometry, the entity model, the A5 import-purity guard, and the level validator (phase 02), plus the tick itself (phase 03) — the seeded RNG's published golden vectors (`test_rng.py`), the kernel's clamping/accumulation/purity/determinism suite (`test_step_kernel.py`), and the movement stage: hand-computed geometry per rule (`test_movement.py`), a scripted 2000-substep four-cook walk on the sim's own RNG that re-checks containment, non-overlap, bounded displacement and hash equality after every substep (`test_movement_stress.py`), and the same walk through the whole `step()` on the shipped levels (`test_movement_kernel.py`), and the action button (phase 03 task 03): the reach-zone geometry and one-press-one-event press edge (`test_targeting.py`), and the table-driven verb ladder that covers every `Verb` and every `DeniedReason` plus the gate's contested arbitration (`test_verbs.py`); the golden-state-hash replay suites — `test_replay_carry.py` (task 04's two-player pass and contest), `test_replay.py` (the committed goldens against `sim/replay.py`'s input-script format, the `record`/`check` CLI and the perf seam) — and the `slow`-marked perf smoke (`test_perf_smoke.py`) | none — the sim must stay callable without a server |
 | `tests/integration/` | multi-module, in-process tests: app boot through `create_app()`, HTTP routes, static serving, room/protocol behaviour with real WebSocket clients (room + protocol integration is PLAN §9 layer 4, and lives here) | in-process app, loopback sockets |
 | `tests/e2e/` | one Playwright suite per user story, named `test_usNN_*.py`; multiplayer stories drive **two browser contexts** against one room | real browser, real socket |
-| `tests/fixtures/` | fixture data only. `tests/fixtures/levels/` is the content dir every test sees — an autouse session fixture points `UNDERCOOKED_LEVELS_DIR` at it so no test can silently read the shipped `levels/` | none |
+| `tests/fixtures/` | fixture data only. `tests/fixtures/levels/` is the content dir every test sees — an autouse session fixture points `UNDERCOOKED_LEVELS_DIR` at it so no test can silently read the shipped `levels/` (`_bad/` holds the deliberately-broken campaign the `_` discovery rule ignores; `_replay/` holds the `pick_drop_pass` fixture kitchen, ignored the same way). `tests/fixtures/replays/` holds the committed input scripts and `tests/goldens/` the committed golden hashes they are gated on (`sim/replay.py`) | none |
 
 There is **no `tests/server/`**: PLAN §9 counts server/route layers of coverage, not
 directories. A route asserted as a pure function (header values, status codes, CLI