test: address AI findings in recent test changes

[mldangelo-oai]

Summary

• clarify several recent test helpers and thresholds surfaced by GitHub AI findings
• make the NumPy v2 coverage test write and assert a real v2 .npy file
• make large-file fallback coverage use realistic scanner fakes while preserving the intended fail-closed behavior

Validation

• UV_CACHE_DIR=/tmp/modelaudit-uv-cache /opt/homebrew/bin/uv run ruff format modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
• UV_CACHE_DIR=/tmp/modelaudit-uv-cache /opt/homebrew/bin/uv run ruff check --fix modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
• UV_CACHE_DIR=/tmp/modelaudit-uv-cache /opt/homebrew/bin/uv run mypy modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
• UV_CACHE_DIR=/tmp/modelaudit-uv-cache PROMPTFOO_DISABLE_TELEMETRY=1 /opt/homebrew/bin/uv run pytest -n auto -m "not slow and not integration" --maxfail=1

[github-actions]

Workflow run and artifacts

Performance Benchmarks

Compared 12 shared benchmarks with a regression threshold of 15%.
Status: 0 regressions, 0 improved, 12 stable, 0 new, 0 missing.
Aggregate shared-benchmark median: 622.05ms -> 627.02ms (+0.8%).

Workload	Benchmark	Target	Size	Files	Baseline	Current	Change	Status
suspicious-pickle-intake	tests/benchmarks/test_scan_benchmarks.py::test_scan_suspicious_pickle_intake	suspicious-intake	183.8 KiB	4	72.92ms	78.99ms	+8.3%	stable
clean-training-checkpoint	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_clean_training_checkpoint	safe_large	278.2 KiB	1	16.62ms	15.48ms	-6.8%	stable
nested-payload-review	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_base64]	nested_base64	98 B	1	133.9us	131.2us	-2.1%	stable
chunked-upload-stream	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_chunked_upload_stream	chunked_stream	278.2 KiB	1	18.74ms	18.40ms	-1.8%	stable
padded-multi-stream-upload	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_padded_multi_stream_upload	multi_stream_padded	4.1 KiB	1	460.5us	466.5us	+1.3%	stable
single-checkpoint-preflight	tests/benchmarks/test_scan_benchmarks.py::test_scan_single_checkpoint_before_load	single_checkpoint.pkl	183.0 KiB	1	35.85ms	35.50ms	-1.0%	stable
nested-payload-review	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_hex]	nested_hex	130 B	1	136.7us	135.4us	-0.9%	stable
mixed-model-repository	tests/benchmarks/test_scan_benchmarks.py::test_scan_release_candidate_repository	release-candidate	547.3 KiB	32	250.17ms	251.59ms	+0.6%	stable
direct-malicious-upload	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_direct_malicious_upload	malicious_reduce	52 B	1	402.7us	400.9us	-0.4%	stable
duplicate-heavy-registry	tests/benchmarks/test_scan_benchmarks.py::test_scan_duplicate_registry_snapshot	registry-snapshot	915.2 KiB	13	190.05ms	189.35ms	-0.4%	stable
nested-payload-review	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_raw]	nested_raw	78 B	1	126.8us	126.3us	-0.3%	stable
warm-cache-rescan	tests/benchmarks/test_scan_benchmarks.py::test_scan_warm_cached_repository_rescan	release-candidate	547.3 KiB	32	36.45ms	36.44ms	-0.0%	stable