Fix Kamino contact capacity allocation

.github/workflows/aws_gpu_benchmarks.yml

@@ -112,6 +112,7 @@ jobs:
         with:
           ref: ${{ inputs.ref }}
           fetch-depth: 0
+          persist-credentials: false
 
       - name: Install system dependencies
         run: |
@@ -137,12 +138,16 @@ jobs:
 
       - name: Run Benchmarks
         run: |
+          # Gate this PR job on directly measured runtimes only.
           uvx --with virtualenv asv continuous \
             --launch-method spawn \
             --interleave-rounds \
             --append-samples \
             --no-only-changed \
-            -e -b Fast \
+            -e \
+            -b 'Fast\w*\.time_' \
+            -b 'Fast(?!NewtonOverhead)\w*\.track_simulate(\(|$)' \
+            -b 'Fast\w*\.track_(mean|p95)_\w*(time|ms)(\(|$)' \
             ${{ inputs.base_ref }} \
             ${{ inputs.ref }}
         continue-on-error: true

.github/workflows/pr_auto_assign_creator.yml

@@ -49,11 +49,16 @@ jobs:
             }
 
             try {
-              await github.rest.issues.checkUserCanBeAssigned({
-                owner,
-                repo,
-                assignee: author,
-              });
+              // PR participants may be assignable without repository-wide access.
+              await github.request(
+                'GET /repos/{owner}/{repo}/issues/{issue_number}/assignees/{assignee}',
+                {
+                  owner,
+                  repo,
+                  issue_number: pr.number,
+                  assignee: author,
+                }
+              );
             } catch (error) {
               if (error.status === 404) {
                 core.warning(

.github/workflows/pr_target_aws_gpu_benchmarks.yml

@@ -96,7 +96,17 @@ jobs:
     needs:
       - check-author-membership
       - require-approval
-    if: github.repository == 'newton-physics/newton' && (!cancelled())
+    if: >-
+      github.repository == 'newton-physics/newton' &&
+      !cancelled() &&
+      needs.check-author-membership.result == 'success' &&
+      (
+        needs.require-approval.result == 'success' ||
+        (
+          needs.require-approval.result == 'skipped' &&
+          needs.check-author-membership.outputs.membership_status == 'CONFIRMED_MEMBER'
+        )
+      )
     uses: ./.github/workflows/aws_gpu_benchmarks.yml
     with:
       ref: ${{ github.event.pull_request.head.sha }}

.github/workflows/pr_target_aws_gpu_tests.yml

@@ -96,7 +96,17 @@ jobs:
     needs:
       - check-author-membership
       - require-approval
-    if: github.repository == 'newton-physics/newton' && (!cancelled())
+    if: >-
+      github.repository == 'newton-physics/newton' &&
+      !cancelled() &&
+      needs.check-author-membership.result == 'success' &&
+      (
+        needs.require-approval.result == 'success' ||
+        (
+          needs.require-approval.result == 'skipped' &&
+          needs.check-author-membership.outputs.membership_status == 'CONFIRMED_MEMBER'
+        )
+      )
     uses: ./.github/workflows/aws_gpu_tests.yml
     with:
       ref: ${{ github.event.pull_request.head.sha }}

.github/workflows/release.yml

@@ -15,7 +15,6 @@ jobs:
     runs-on: ubuntu-latest
     permissions:
       contents: read
-      actions: write
     steps:
       - name: Harden the runner (Audit all outbound calls)
         uses: step-security/harden-runner@fa2e9d605c4eeb9fcad4c99c224cee0c6c7f3594  # v2.16.0

AGENTS.md

@@ -40,6 +40,7 @@ uv run --extra dev --extra torch-cu12 -m newton.tests                  # with Py
 
 ### Testing guidelines
 
+- Give every test function or method a docstring using triple double quotes (`"""..."""`). Start with a concise one-line summary in imperative mood that states what the test verifies. For a particularly complex test, add a body that elaborates on the tested behavior, separated from the summary by a blank line following Google-style docstring conventions.
 - Never call `wp.synchronize()` or `wp.synchronize_device()` right before `.numpy()` on a Warp array. This is redundant as `.numpy()` performs a synchronous device-to-host copy that completes all outstanding work.
 
 ```bash

README.md

@@ -283,6 +283,9 @@ If you run the examples from a source checkout with uv, use
       </a>
     </td>
     <td align="center" width="33%">
+      <a href="https://github.com/newton-physics/newton/blob/main/newton/examples/cable/example_cable_plectoneme.py">
+        <img width="320" src="https://raw.githubusercontent.com/newton-physics/newton/main/docs/images/examples/example_cable_plectoneme.jpg" alt="Cable Plectoneme">
+      </a>
     </td>
   </tr>
   <tr>
@@ -293,6 +296,7 @@ If you run the examples from a source checkout with uv, use
       <code>python -m newton.examples cable_cross_slide_table</code>
     </td>
     <td align="center" width="33%">
+      <code>python -m newton.examples cable_plectoneme</code>
     </td>
   </tr>
   <tr>
@@ -749,6 +753,9 @@ If you run the examples from a source checkout with uv, use
       </a>
     </td>
     <td align="center" width="33%">
+      <a href="https://github.com/newton-physics/newton/blob/main/newton/examples/contacts/example_newton_cradle.py">
+        <img width="320" src="https://raw.githubusercontent.com/newton-physics/newton/main/docs/images/examples/example_newton_cradle.jpg" alt="Newton Cradle">
+      </a>
     </td>
   </tr>
   <tr>
@@ -758,6 +765,31 @@ If you run the examples from a source checkout with uv, use
     <td align="center" width="33%">
       <code>python -m newton.examples contacts_rj45_plug</code>
     </td>
+    <td align="center" width="33%">
+      <code>python -m newton.examples newton_cradle</code>
+    </td>
+  </tr>
+  <tr>
+    <td align="center" width="33%">
+      <a href="https://github.com/newton-physics/newton/blob/main/newton/examples/contacts/example_balance_bird.py">
+        <img width="320" src="https://raw.githubusercontent.com/newton-physics/newton/main/docs/images/examples/example_balance_bird.jpg" alt="Balance Bird">
+      </a>
+    </td>
+    <td align="center" width="33%">
+      <a href="https://github.com/newton-physics/newton/blob/main/newton/examples/contacts/example_domino_spiral.py">
+        <img width="320" src="https://raw.githubusercontent.com/newton-physics/newton/main/docs/images/examples/example_domino_spiral.jpg" alt="Domino Spiral">
+      </a>
+    </td>
+    <td align="center" width="33%">
+    </td>
+  </tr>
+  <tr>
+    <td align="center" width="33%">
+      <code>python -m newton.examples balance_bird</code>
+    </td>
+    <td align="center" width="33%">
+      <code>python -m newton.examples domino_spiral</code>
+    </td>
     <td align="center" width="33%">
     </td>
   </tr>

asv.conf.json

@@ -16,7 +16,7 @@
   "build_command": ["python -m build --wheel -o {build_cache_dir} {build_dir}"],
   "install_command": [
     "in-dir={env_dir} python -m pip install {wheel_file}[dev] --extra-index-url=https://pypi.nvidia.com/",
-    "in-dir={env_dir} python -m pip install {wheel_file}[sim] warp-lang==1.16.0.dev20260716 mujoco==3.10.0 mujoco-warp==3.10.0.2 --extra-index-url=https://pypi.nvidia.com/",
+    "in-dir={env_dir} python -m pip install {wheel_file}[sim] warp-lang==1.16.0.dev20260716 mujoco==3.10.0 mujoco-warp==3.10.0.3 --extra-index-url=https://pypi.nvidia.com/",
     "python -m pip install torch==2.10.0+cu130 --index-url https://download.pytorch.org/whl/cu130"
   ]
 }

asv/benchmarks/benchmark_kamino.py

@@ -12,6 +12,11 @@
 
 import newton
 
+if __package__:
+    from .benchmark_metrics import validate_simulation_state
+else:
+    from benchmark_metrics import validate_simulation_state
+
 _NUM_ACTIONS = 12
 _OBS_DIM = 94
 _MIN_STANDING_HEIGHT = 0.20
@@ -435,37 +440,23 @@ def step(self):
         self.sim_time += self.frame_dt
 
     def test_final(self):
-        state_values = {}
-        for name in ("joint_q", "body_q", "body_qd"):
-            values = getattr(self.state_0, name).numpy()
-            if not np.isfinite(values).all():
-                raise RuntimeError(f"Simulation produced non-finite values in state.{name}")
-            state_values[name] = values
-
-        body_count = self.model.body_count // self.world_count
-        body_qd = state_values["body_qd"].reshape(self.world_count, body_count, 6)
-        max_linear_speed = np.linalg.norm(body_qd[:, :, :3], axis=-1).max()
-        max_angular_speed = np.linalg.norm(body_qd[:, :, 3:], axis=-1).max()
-        if max_linear_speed > _MAX_BODY_LINEAR_SPEED:
-            raise RuntimeError(
-                f"Maximum body linear speed is {max_linear_speed:.3f} m/s, exceeding {_MAX_BODY_LINEAR_SPEED:.1f} m/s"
-            )
-        if max_angular_speed > _MAX_BODY_ANGULAR_SPEED:
-            raise RuntimeError(
-                f"Maximum body angular speed is {max_angular_speed:.3f} rad/s, "
-                f"exceeding {_MAX_BODY_ANGULAR_SPEED:.1f} rad/s"
-            )
+        validate_simulation_state(
+            self.state_0,
+            max_linear_speed=_MAX_BODY_LINEAR_SPEED,
+            max_angular_speed=_MAX_BODY_ANGULAR_SPEED,
+        )
 
         if self.policy_controller is None:
             return
 
+        body_count = self.model.body_count // self.world_count
         body_labels = [label.rsplit("/", 1)[-1] for label in self.model.body_label[:body_count]]
         try:
             pelvis_index = body_labels.index("pelvis")
         except ValueError as e:
             raise RuntimeError("DR Legs model has no pelvis root body") from e
 
-        body_q = state_values["body_q"].reshape(self.world_count, body_count, 7)[:, pelvis_index]
+        body_q = self.state_0.body_q.numpy().reshape(self.world_count, body_count, 7)[:, pelvis_index]
         body_com = self.model.body_com.numpy().reshape(self.world_count, body_count, 3)[:, pelvis_index]
         quat_vector = body_q[:, 3:6]
         twice_cross = 2.0 * np.cross(quat_vector, body_com)