feat(cccl): integrate missing CCCL directories — python/, ci/, .agent/, docs/, test/

Sparse-checkout from NVIDIA/cccl main branch to complete cccl_upstream:

Added:
- python/cuda_cccl/ (226 files) — Python bindings for device-level algorithms
  Critical for muh toolchain: cuda.compute.reduce_into, scan, radix_sort, etc.
  Includes 204 .py files with full test coverage for all 27 algorithms
- ci/ (163 files) — Build/test infrastructure
  build_cub.sh, test_cub.sh, build_and_test_targets.sh, matrix.yaml
  Directly maps to our [INFRA-CI] and [INFRA-BUILD] items
- .agent/skills/ (7 files) — NVIDIA's own agent skills for CCCL
  cccl-style/SKILL.md, cccl-test/SKILL.md, sass-diff/SKILL.md
- docs/ (491 files) — Official CCCL documentation
  CI references, CMake guides, Python compute docs, libcudacxx PTX docs
- test/ (12 files) — Top-level integration tests (cuda_smoke, stdpar)
- Root configs: .clang-format, .clang-tidy, CONTRIBUTING.md, pyproject.toml
- CLAUDE.md symlink → AGENTS.md (NVIDIA's standard)

cccl_upstream now mirrors full NVIDIA/cccl structure:
  Before: 42M (cub + thrust + libcudacxx + cudax + c + examples + benchmarks)
  After:  53M (+python +ci +docs +.agent +test +configs)

This completes the CCCL base needed for:
- [muh-bench] items: ci/util/build_and_test_targets.sh for targeted builds
- [CCCL-verify] items: python/cuda_cccl/tests/ as reference implementations
- [CCCL-test] items: ci/test_cub.sh, ci/test_thrust.sh
- Agent workflow: .agent/skills/ for consistent style and test patterns

cccl_upstream/docs/libcudacxx/extended_api/mdspan.rst

@@ -0,0 +1,54 @@
+.. _libcudacxx-extended-api-mdspan:
+
+Mdspan
+======
+
+.. toctree::
+   :hidden:
+   :maxdepth: 1
+
+   mdspan/host_device_accessor
+   mdspan/layout_stride_relaxed
+   mdspan/restrict_accessor
+   mdspan/shared_memory_accessor
+   mdspan/mdspan_to_dlpack
+   mdspan/dlpack_to_mdspan
+
+.. list-table::
+   :widths: 25 45 30 30
+   :header-rows: 1
+
+   * - **Header**
+     - **Content**
+     - **CCCL Availability**
+     - **CUDA Toolkit Availability**
+
+   * - :ref:`host/device/managed mdspan and accessor <libcudacxx-extended-api-mdspan-host-device-accessor>`
+     - CUDA memory space ``mdspan`` and accessors
+     - CCCL 3.0.0
+     - CUDA 13.0
+
+   * - :ref:`layout_stride_relaxed <libcudacxx-extended-api-mdspan-layout-stride-relaxed>`
+     - Layout mapping policy with negative/zero strides and offset support
+     - CCCL 3.0.0
+     - CUDA 13.0
+
+   * - :ref:`restrict mdspan and accessor <libcudacxx-extended-api-mdspan-restrict-accessor>`
+     - ``mdspan`` and accessor with the *restrict* aliasing policy
+     - CCCL 3.0.0
+     - CUDA 13.0
+
+   * - :ref:`shared_memory mdspan and accessor <libcudacxx-extended-api-mdspan-shared-memory-accessor>`
+     - ``mdspan`` and accessor for CUDA shared memory
+     - CCCL 3.3.0
+     - CUDA 13.3
+
+   * - :ref:`mdspan to dlpack <libcudacxx-extended-api-mdspan-mdspan-to-dlpack>`
+     - Convert a ``mdspan`` to a ``DLTensor``
+     - CCCL 3.2.0
+     - CUDA 13.2
+
+   * - :ref:`dlpack to mdspan <libcudacxx-extended-api-mdspan-dlpack-to-mdspan>`
+     - Convert a ``DLTensor`` to a ``mdspan``
+     - CCCL 3.2.0
+     - CUDA 13.2