Verification oracle: tests/oracle cases, cargo xtask oracle, fixtures, docs/oracle.md, CI job

The oracle binary (modelcraft-testkit) runs each case's recipe through the
command engine, measures its quantities by JSON pointers into step outputs
or commands, compares them with the cited references and the frozen
fixtures, and writes docs/oracle.md. 20 theory cases (patch tests of every
solid element, Timoshenko cantilevers, Lame, Kirsch, beam frequencies,
torsion and bending through couplings, explicit waves and impact, CDT
invariants), 8 published benchmarks (MacNeal-Harder patch, straight and
curved cantilevers, thick cylinder, Scordelis-Lo with solids, NAFEMS LE1,
LE10, FV32) and 10 pending cases with citations.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NmAttHhUZTokgPeEJia8sm

tests/oracle/pure-bending-coupling/expected.json

@@ -0,0 +1,44 @@
+{
+ "title": "Pure bending through a distributing coupling",
+ "kind": "theory",
+ "source": "Euler–Bernoulli pure bending, σ = M_y z/I, θ = ML/(EI) (exact in elasticity for ν = 0); Timoshenko and Goodier, Theory of Elasticity, 3rd ed., 1970, §102",
+ "quantities": [
+  {
+   "id": "sigma_top",
+   "quantity": "bending stress σ_xx at mid-span, top fibre (10, 0.5, 1)",
+   "measure": {
+    "command": "solve.probe",
+    "params": {
+     "step": "Bend",
+     "at": [
+      [
+       10.0,
+       0.5,
+       1.0
+      ]
+     ]
+    },
+    "pointer": "/nodes/0/stress/0"
+   },
+   "value": 15.0,
+   "tolerance": {
+    "relative": 0.005
+   },
+   "reason": "with ν = 0 pure bending is a quadratic displacement field that hex20 represents exactly; the coupling's nodal force pattern differs from the linear traction only within a depth or two of the end (Saint-Venant), five depths from the probe"
+  },
+  {
+   "id": "rotation",
+   "quantity": "rotation of the reference point about y",
+   "measure": {
+    "from": "solve",
+    "pointer": "/reference_points/0/displacement/4"
+   },
+   "value": 0.30000000000000004,
+   "tolerance": {
+    "relative": 0.01
+   },
+   "reason": "the reference point follows the best-fit rigid motion of the end section, which the end zone's local deformation disturbs by well under 1%",
+   "known_failure": "1.2% high: the distributing coupling weights every node of the end face equally, corner and mid-side nodes of hex20 alike, which is not the consistent linear traction; the end face deforms locally and the fitted rotation includes it. Area-consistent weights would remove it"
+  }
+ ]
+}