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/ptx/instructions/manual/shfl_sync.rst

@@ -0,0 +1,80 @@
+
+shfl.sync
+^^^^^^^^^
+
+.. code:: cuda
+
+   // PTX ISA 6.0
+   // shfl.sync.mode.b32  d[|p], a, b, c, membermask;
+   //   .mode = { .up, .down, .bfly, .idx };
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_idx(T        data,
+                   uint32_t lane_idx_offset,
+                   uint32_t clamp_segmask,
+                   uint32_t lane_mask) noexcept;
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_idx(T        data,
+                   bool&    pred,
+                   uint32_t lane_idx_offset,
+                   uint32_t clamp_segmask,
+                   uint32_t lane_mask) noexcept;
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_up(T        data,
+                  uint32_t lane_idx_offset,
+                  uint32_t clamp_segmask,
+                  uint32_t lane_mask) noexcept;
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_up(T        data,
+                  bool&    pred,
+                  uint32_t lane_idx_offset,
+                  uint32_t clamp_segmask,
+                  uint32_t lane_mask) noexcept;
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_down(T        data,
+                    uint32_t lane_idx_offset,
+                    uint32_t clamp_segmask,
+                    uint32_t lane_mask) noexcept;
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_down(T        data,
+                    bool&    pred,
+                    uint32_t lane_idx_offset,
+                    uint32_t clamp_segmask,
+                    uint32_t lane_mask) noexcept;
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_bfly(T        data,
+                    uint32_t lane_idx_offset,
+                    uint32_t clamp_segmask,
+                    uint32_t lane_mask) noexcept;
+
+   template<typename T>
+   [[nodiscard]] __device__ static inline
+   T shfl_sync_bfly(T        data,
+                    bool&    pred,
+                    uint32_t lane_idx_offset,
+                    uint32_t clamp_segmask,
+                    uint32_t lane_mask) noexcept;
+
+**Constrains and checks**
+
+The following conditions are checked at *run-time* in debug mode:
+
+- ``T`` must have 32-bit size (compile-time)
+- ``lane_idx_offset`` must be less than the warp size
+- ``clamp_segmask`` must use the bit positions [0:4] and [8:12])
+- ``lane_mask`` must be a subset of the active mask
+- All lanes must have the same value for ``lane_mask``
+- The destination lane must be a member of the ``lane_mask``