fix: fail closed on executable ZIP scanner gaps

[mldangelo-oai]

Summary

• fail closed when an executable ZIP composed subtype is recognized but its scanner cannot load
• keep generic ZIP member scanning while preserving the operational incomplete result
• add a Skops-shaped composed ZIP regression plus existing polyglot guards

Validation

• PYTHONPATH=/private/tmp/modelaudit-c125 PROMPTFOO_DISABLE_TELEMETRY=1 /Users/mdangelo/code/modelaudit/.venv/bin/pytest tests/scanners/test_zip_scanner.py::test_executable_zip_composed_routing_fails_closed_when_subtype_scanner_unavailable tests/scanners/test_zip_scanner.py::test_scan_nested_file_fails_closed_when_recognized_header_scanner_is_unavailable tests/test_directory_file_filtering.py::TestDirectoryFileFiltering::test_disguised_llamafile_skops_polyglot_preserves_cve_findings tests/test_directory_file_filtering.py::TestDirectoryFileFiltering::test_out_of_window_executable_skops_zip_preserves_cve_findings -q
• ruff format --check modelaudit/scanners/archive_dispatch.py tests/scanners/test_zip_scanner.py
• ruff check modelaudit/scanners/archive_dispatch.py tests/scanners/test_zip_scanner.py
• mypy modelaudit/scanners/archive_dispatch.py tests/scanners/test_zip_scanner.py
• git diff --check

[mldangelo-oai]

@codex review

[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: 1.277s -> 1.284s (+0.5%).

Workload	Benchmark	Target	Size	Files	Baseline	Current	Change	Status
nested-payload-review	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_raw]	nested_raw	78 B	1	416.4us	396.9us	-4.7%	stable
nested-payload-review	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_base64]	nested_base64	98 B	1	422.9us	405.9us	-4.0%	stable
nested-payload-review	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_hex]	nested_hex	130 B	1	432.4us	417.6us	-3.4%	stable
padded-multi-stream-upload	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_padded_multi_stream_upload	multi_stream_padded	4.1 KiB	1	1.53ms	1.50ms	-2.3%	stable
direct-malicious-upload	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_direct_malicious_upload	malicious_reduce	52 B	1	1.45ms	1.42ms	-2.0%	stable
single-checkpoint-preflight	tests/benchmarks/test_scan_benchmarks.py::test_scan_single_checkpoint_before_load	single_checkpoint.pkl	183.0 KiB	1	70.83ms	72.21ms	+1.9%	stable
warm-cache-rescan	tests/benchmarks/test_scan_benchmarks.py::test_scan_warm_cached_repository_rescan	release-candidate	547.3 KiB	32	65.33ms	64.49ms	-1.3%	stable
clean-training-checkpoint	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_clean_training_checkpoint	safe_large	278.2 KiB	1	110.18ms	111.55ms	+1.2%	stable
mixed-model-repository	tests/benchmarks/test_scan_benchmarks.py::test_scan_release_candidate_repository	release-candidate	547.3 KiB	32	402.35ms	406.92ms	+1.1%	stable
chunked-upload-stream	tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_chunked_upload_stream	chunked_stream	278.2 KiB	1	113.53ms	114.29ms	+0.7%	stable
suspicious-pickle-intake	tests/benchmarks/test_scan_benchmarks.py::test_scan_suspicious_pickle_intake	suspicious-intake	183.8 KiB	4	131.15ms	130.94ms	-0.2%	stable
duplicate-heavy-registry	tests/benchmarks/test_scan_benchmarks.py::test_scan_duplicate_registry_snapshot	registry-snapshot	915.2 KiB	13	379.59ms	379.18ms	-0.1%	stable

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. 👍

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[ianw-oai]

Looks good.