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/ci/upload_thrust_test_artifacts.sh

@@ -0,0 +1,93 @@
+#!/usr/bin/env bash
+
+set -euo pipefail
+
+if [[ -z "${GITHUB_ACTIONS:-}" ]]; then
+  echo "This script must be run in a GitHub Actions environment." >&2
+  exit 1
+fi
+
+ci_dir="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+readonly ci_dir
+repo_root="$(cd "${ci_dir}/.." && pwd)"
+readonly repo_root
+
+cd "$repo_root"
+
+if ! ci/util/workflow/has_consumers.sh; then
+  echo "No consumers found for this job. Exiting." >&2
+  exit 0
+fi
+
+# Figure out which artifact needs to be built:
+consumers=$(ci/util/workflow/get_consumers.sh)
+preset_variants=()
+if grep -q "TestGPU" <<< "$consumers"; then
+  preset_variants+=("test_gpu")
+fi
+if grep -q "TestCPU" <<< "$consumers"; then
+  preset_variants+=("test_cpu")
+fi
+
+artifact_prefix="z_thrust-test-artifacts-${DEVCONTAINER_NAME:?}-${JOB_ID:?}"
+
+# BUILD_INFIX is undefined on windows CI
+build_dir_regex="build${CCCL_BUILD_INFIX:+/$CCCL_BUILD_INFIX}/thrust[^/]*"
+
+# Just collect the minimum set of files needed for running each ctest preset:
+for preset_variant in "${preset_variants[@]}"; do
+  # Shared across all presets:
+  ci/util/artifacts/stage.sh "$artifact_prefix-$preset_variant" \
+      "$build_dir_regex/build\.ninja$" \
+      "$build_dir_regex/.*rules\.ninja$" \
+      "$build_dir_regex/CMakeCache\.txt$" \
+      "$build_dir_regex/.*VerifyGlobs\.cmake$" \
+      "$build_dir_regex/.*CTestTestfile\.cmake$" \
+      > /dev/null
+done
+
+if [[ " ${preset_variants[*]} " == *" test_cpu "* ]]; then
+  # Initially add all binaries, then remove all containing 'cuda' in the name:
+  ci/util/artifacts/stage.sh \
+      "$artifact_prefix-test_cpu" \
+      "$build_dir_regex/bin/.*" > /dev/null
+  ci/util/artifacts/unstage.sh \
+      "$artifact_prefix-test_cpu" \
+      "$build_dir_regex/bin/thrust\..*\.cuda\..*" > /dev/null
+
+  ci/util/artifacts/stage.sh \
+      "$artifact_prefix-test_cpu" \
+      "$build_dir_regex/lib/.*\.test\.framework\..*" > /dev/null
+  ci/util/artifacts/unstage.sh \
+      "$artifact_prefix-test_cpu" \
+      "$build_dir_regex/lib/.*\.cuda\.test\.framework\..*" > /dev/null
+
+  # Windows builds generate binaries for the header tests, remove these:
+  ci/util/artifacts/unstage.sh  \
+      "$artifact_prefix-test_cpu" \
+      "$build_dir_regex/.*\.headers\..*" > /dev/null || :
+
+  ci/util/artifacts/upload_stage_packed.sh "$artifact_prefix-test_cpu"
+fi
+
+if [[ " ${preset_variants[*]} " == *" test_gpu "* ]]; then
+  # Only binaries containing 'cuda':
+  ci/util/artifacts/stage.sh \
+      "$artifact_prefix-test_gpu" \
+      "$build_dir_regex/bin/thrust\..*\.cuda\..*" > /dev/null
+  ci/util/artifacts/stage.sh \
+      "$artifact_prefix-test_gpu" \
+      "$build_dir_regex/lib/.*\.cuda\.test\.framework\..*" > /dev/null
+  # The CUDA runtime smoke binary is invoked explicitly from build_common.sh
+  ci/util/artifacts/stage.sh \
+      "$artifact_prefix-test_gpu" \
+      "$build_dir_regex/bin/cccl\.test\.cuda_runtime_smoke$" > /dev/null
+
+  # Windows builds generate binaries for the header tests, remove these:
+  ci/util/artifacts/unstage.sh  \
+      "$artifact_prefix-test_gpu" \
+      "$build_dir_regex/.*\.headers\..*" > /dev/null || :
+
+
+  ci/util/artifacts/upload_stage_packed.sh "$artifact_prefix-test_gpu"
+fi