fix: preserve ONNX external data in HF streaming by mldangelo-oai · Pull Request #1635 · promptfoo/modelaudit

mldangelo-oai · 2026-06-11T00:50:38Z

Summary

Preserve safe, declared ONNX external_data companions during Hugging Face streaming downloads.
Download sidecars only after parsing the parent ONNX with load_external_data=False, require repo-listing membership, and keep them beside the parent until the parent scan completes.
Keep unsafe, missing, or oversized sidecars fail-closed/visible through existing scanner checks and download budgets.

Root Cause

HF streaming selected files by scanner-owned suffix/content routing. BAAI/bge-m3 declares onnx/model.onnx_data from onnx/model.onnx, but .onnx_data is not a scanner extension, so streaming downloaded only model.onnx. The ONNX scanner then resolved the declared sidecar relative to the streamed temp path and reported a false missing-external-data warning. Local scans did not warn because the sibling file was present.

Security Tradeoffs

Traversal, Windows absolute, backslash, missing, and non-repo-listed external paths are not downloaded or normalized into safe-looking files; OnnxScanner still sees the original model metadata and reports the existing CVE/missing checks.
Sidecar sizes are fetched from the pinned revision and counted against max_size before download, then verified after materialization.
Context sidecars are not yielded as standalone model scans; they are cleaned up when the generator advances/closes after the parent scan.

Pinned BGE-M3 QA

Pinned revision: BAAI/bge-m3@5617a9f61b028005a4858fdac845db406aefb181.

Pre-fix confirmation on the pinned main SHA 2f782ba1f18ab5aac4716c8c7ac6f0a16f0c6f60:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run python <pinned streaming reproduction>

Outcome: streaming scanned only onnx/model.onnx (724923 bytes), exited 1, and emitted External data reference found (file may not be present): 'model.onnx_data' (389 tensors affected).

Post-fix live test:

MODELAUDIT_RUN_REAL_HF_BGE_M3=1 PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming::test_download_model_streaming_real_bge_m3_onnx_external_data_pinned -q

Outcome: 1 passed in 14.34s.

Post-fix streaming/local comparison script outcome:

revision 5617a9f61b028005a4858fdac845db406aefb181
sidecar_bytes 2266820608
stream_exit 2 stream_success False stream_files 1 stream_bytes 724923
stream_missing_resolved_size_weight (0, 1, 389, 1)
local_success False local_checks 400 local_bytes 724923
local_missing_resolved_size_weight (0, 1, 389, 1)

The remaining exit 2 is the existing ONNX weight-distribution coverage outcome and now matches local behavior; the missing-sidecar warning is gone.

Validation

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming -k 'onnx_external_data or bge_m3' -q
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/test_streaming_scan.py -k 'hf_onnx' -q
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/utils/sources/test_huggingface.py tests/test_streaming_scan.py -q --maxfail=1
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_onnx_scanner.py -k 'external_data' -q --maxfail=1
uv run ruff check modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
uv run ruff format --check modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
uv run mypy modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1
git diff --check

Outcomes: focused suites passed; full mypy passed (474 source files); full non-slow/non-integration pytest passed (18065 passed, 793 skipped, 31 warnings in 780.49s); Ruff and diff checks passed.

Follow-up Validation (`4ed5fd2`)

Resolved the three follow-up P2 threads:

Selected ONNX sidecars that appear before the parent are tracked in downloaded_selected_paths, so they are not re-downloaded or counted twice under an exact max_size budget.
HF cache .onnx snapshot aliases now keep scan identity by snapshot context, while ordinary non-sharded HF cache aliases still scan and dedupe by blob path.
Trusted ONNX external-data symlink exceptions now require lexical snapshots/<revision>/... aliases for the same cache root/revision; arbitrary directories under models--* retain CVE-2026-34447 checks.

Pinned XLM-R streaming/local QA:

repo FacebookAI/xlm-roberta-large@c23d21b0620b635a76227c604d44e43a9f0ee389
model_size 545850
sidecar_size 2235363328
exact_max_size 2235909178
stream_summary {'success': False, 'exit': 2, 'checks': 400, 'issues': 1, 'bytes_scanned': 545850, 'files_scanned': 1, 'failed_external': 0, 'passed_external': 1, 'size_failures': 0, 'cves': []}
stream_cache_remaining_bytes 442
local_summary {'success': False, 'exit': 2, 'checks': 12, 'issues': 1, 'bytes_scanned': 2235909178, 'files_scanned': 2, 'failed_external': 0, 'passed_external': 1, 'size_failures': 0, 'cves': []}
local_download_bytes 2235909620
elapsed_seconds 29.86
ru_maxrss_kb 142528

Additional validation:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/utils/sources/test_huggingface.py tests/test_streaming_scan.py -q --maxfail=1
# 360 passed, 1 skipped
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_onnx_scanner.py -k 'external_data or CVE202634447' -q --maxfail=1
# 16 passed, 233 deselected
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/test_core.py -k 'hf_snapshot or trusted_hf_alias or hf_shard or snapshot_alias' -q --maxfail=1
# 8 passed, 633 deselected
uv run ruff check modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
# All checks passed
uv run ruff format --check modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
# 419 files already formatted
uv run mypy modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
# Success: no issues found in 474 source files
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1
# 18554 passed, 793 skipped, 40 warnings in 780.18s
git diff --check
# clean

Follow-up Validation (`da12fe6`)

Tightened HF cache external-data trust to per-file snapshots/<revision>/... symlink leaves; symlinked parent directories into blobs now retain CVE-2026-34447 coverage.
Re-fetches previously yielded selected ONNX sidecars when delete_after_scan=True removed them before the parent ONNX scan, without double-counting the exact selected-file max-size budget.
Added regressions for deleted selected sidecars and snapshot directory symlinks into cache blobs.
Local validation: targeted new tests, affected HF/ONNX suites, ruff check --fix, ruff format --check, mypy, git diff --check, and PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1 (18556 passed, 793 skipped).

Final-head pinned XLM-R QA (da12fe667b743bad11e09caf5976ad7ad1291b84): FacebookAI/xlm-roberta-large@c23d21b0620b635a76227c604d44e43a9f0ee389, exact budget 2235909178 bytes. Streaming summary: files_scanned=1, bytes_scanned=545850, passed_external=1, failed_external=0, size_failures=0, cves=[], exit=2 only from existing fail-closed weight coverage. Local snapshot summary: files_scanned=2, bytes_scanned=2235909178, passed_external=1, failed_external=0, size_failures=0, cves=[], exit=2 for the same coverage reason. Resource evidence: elapsed_seconds=16.87, ru_maxrss_kb=154384, far below the 2.235 GB sidecar size.

mldangelo-oai · 2026-06-11T00:50:59Z

@codex review

github-actions · 2026-06-11T00:52:56Z

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: 1.477s -> 1.504s (+1.8%).

Workload	Benchmark	Target	Size	Files	Baseline	Current	Change	Status
`warm-cache-rescan`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_warm_cached_repository_rescan`	`release-candidate`	547.3 KiB	32	129.80ms	140.74ms	+8.4%	stable
`nested-payload-review`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_raw]`	`nested_raw`	78 B	1	603.5us	579.1us	-4.1%	stable
`padded-multi-stream-upload`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_padded_multi_stream_upload`	`multi_stream_padded`	4.1 KiB	1	662.9us	639.0us	-3.6%	stable
`mixed-model-repository`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_release_candidate_repository`	`release-candidate`	547.3 KiB	32	495.78ms	511.85ms	+3.2%	stable
`direct-malicious-upload`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_direct_malicious_upload`	`malicious_reduce`	52 B	1	529.9us	514.0us	-3.0%	stable
`nested-payload-review`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_base64]`	`nested_base64`	98 B	1	602.1us	584.1us	-3.0%	stable
`nested-payload-review`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_hex]`	`nested_hex`	130 B	1	632.8us	614.6us	-2.9%	stable
`clean-training-checkpoint`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_clean_training_checkpoint`	`safe_large`	278.2 KiB	1	110.72ms	109.00ms	-1.6%	stable
`duplicate-heavy-registry`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_duplicate_registry_snapshot`	`registry-snapshot`	915.2 KiB	13	406.11ms	408.54ms	+0.6%	stable
`single-checkpoint-preflight`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_single_checkpoint_before_load`	`single_checkpoint.pkl`	183.0 KiB	1	71.30ms	71.71ms	+0.6%	stable
`suspicious-pickle-intake`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_suspicious_pickle_intake`	`suspicious-intake`	183.8 KiB	4	146.42ms	145.69ms	-0.5%	stable
`chunked-upload-stream`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_chunked_upload_stream`	`chunked_stream`	278.2 KiB	1	113.60ms	113.37ms	-0.2%	stable

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b7e346ab85

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".

…t08-onnx-external-data-streaming-20260610

mldangelo-oai · 2026-06-11T01:01:50Z

@codex review

chatgpt-codex-connector · 2026-06-11T01:04:52Z

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".

mldangelo-oai · 2026-06-11T01:40:24Z

@codex review

chatgpt-codex-connector · 2026-06-11T01:44:18Z

Codex Review: Didn't find any major issues. You're on a roll.

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".

mldangelo-oai · 2026-06-11T02:26:10Z

Additional pinned QA: FacebookAI/xlm-roberta-large@c23d21b0620b635a76227c604d44e43a9f0ee389 includes onnx/model.onnx plus a declared 2.235 GB onnx/model.onnx_data, but origin/main reports the sidecar as possibly absent. Please run streaming and local-directory scans against the PR head and confirm the false missing-sidecar warning is gone while missing, escaping, swapped, and oversized controls still fail closed. Auto-merge is paused pending this proof.

mldangelo-oai · 2026-06-11T02:48:49Z

Pinned XLM-R ONNX external_data QA on exact PR head 7975772a514ffbe534a3f4b2b48e25b535c8a679.

Target:

FacebookAI/xlm-roberta-large@c23d21b0620b635a76227c604d44e43a9f0ee389
Remote metadata confirmed: onnx/model.onnx = 545850 bytes, onnx/model.onnx_data = 2235363328 bytes.

Commands run:

git fetch origin main
git fetch origin mdangelo/codex/hf-fp-t08-onnx-external-data-streaming-20260610
gh pr view 1635 --repo promptfoo/modelaudit --json headRefOid,headRefName,baseRefOid,mergeStateStatus,reviewDecision,url
PROMPTFOO_DISABLE_TELEMETRY=1 uv run python - <<'PY'
# HfApi().repo_info(repo_id, revision=revision, files_metadata=True)
# patch _list_repo_files_with_timeout -> (repo_files, revision, None)
# scan_model_streaming(download_model_streaming(url, cache_dir=None, max_size=2236000000,
#   timeout_seconds=7200, scannable_extensions={'.onnx'}, scannable_scanner_ids={'onnx'}),
#   timeout=7200, delete_after_scan=False, cache_enabled=False, scanners=['onnx'], skip_file_types=False)
PY
PROMPTFOO_DISABLE_TELEMETRY=1 uv run python - <<'PY'
# scan_model_directory_or_file('~/.cache/huggingface/hub/models--FacebookAI--xlm-roberta-large/snapshots/c23d21b0620b635a76227c604d44e43a9f0ee389/onnx',
#   cache_enabled=False, scanners=['onnx'], skip_file_types=False)
PY

Terminal outcomes:

Streaming HF ONNX-only scan: scanner_names=['onnx'], files_scanned=1, bytes_scanned=545850, issue_count=1, failed_check_count=1, exit_code=2.
- External Data Reference Check: resolved model.onnx_data = 1, missing model.onnx_data = 0.
- External Data Size Validation: passed 389, failed 0.
- Only failed check was unrelated fail-closed coverage: Weight Distribution Analysis Coverage / onnx_weight_distribution_analysis_incomplete / partial_initializer_coverage.
- Peak RSS proxy: self 130052 KiB, child 127408 KiB.
Local-directory ONNX-only scan over the cached snapshot .../snapshots/c23d21b0620b635a76227c604d44e43a9f0ee389/onnx: scanner_names=['onnx'], files_scanned=2, bytes_scanned=2235909178, issue_count=1, failed_check_count=1, exit_code=2.
- External Data Reference Check: resolved model.onnx_data = 1, missing model.onnx_data = 0.
- External Data Size Validation: passed 1, failed 0 after aggregation.
- Only failed check was the same unrelated fail-closed weight-distribution coverage check.
- Peak RSS proxy: self 140588 KiB, child 0 KiB.

Bounded-read evidence: the ONNX parent is parsed with load_external_data=False; sidecar validation uses Hugging Face size metadata and filesystem stat() on the downloaded sidecar. The 2.235 GB sidecar was resolved without reading it into Python scanner memory; the RSS numbers above stay far below sidecar size.

Controls preserved:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -q tests/test_streaming_scan.py::test_scan_model_directory_hf_cache_onnx_external_data_uses_snapshot_alias tests/test_streaming_scan.py::test_scan_model_streaming_hf_onnx_external_data_sidecar_matches_local_directory tests/test_streaming_scan.py::test_scan_model_streaming_hf_onnx_missing_external_data_still_warns tests/test_streaming_scan.py::test_scan_model_streaming_hf_onnx_escaping_external_data_remains_cve tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming::test_download_model_streaming_preserves_onnx_external_data_before_parent_yield tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming::test_download_model_streaming_preserves_content_routed_renamed_onnx_external_data tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming::test_download_model_streaming_include_all_counts_selected_onnx_sidecar_once tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming::test_download_model_streaming_default_hf_cache_preserves_onnx_external_data_sidecar tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming::test_download_model_streaming_does_not_fetch_escaping_onnx_external_data tests/utils/sources/test_huggingface.py::TestModelDownloadStreaming::test_download_model_streaming_blocks_oversized_onnx_external_data
# 10 passed
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -q tests/scanners/test_onnx_scanner.py::test_onnx_scanner_external_data_missing tests/scanners/test_onnx_scanner.py::test_onnx_scanner_external_data_exists tests/scanners/test_onnx_scanner.py::TestCVE202551480SavePathTraversal tests/scanners/test_onnx_scanner.py::TestCVE202634447SymlinkTraversal tests/scanners/test_onnx_scanner.py::TestCVE202427318NestedPathTraversal tests/scanners/test_onnx_scanner.py::TestExternalDataSizeValidation
# 44 passed
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -q tests/test_streaming_scan.py::test_scan_model_streaming_keeps_hf_snapshot_symlink_families_separate tests/test_streaming_scan.py::test_cross_directory_shard_reconciliation_rejects_sequential_hardlinks tests/test_streaming_scan.py::test_scan_model_streaming_does_not_reconcile_duplicate_shard_targets tests/test_streaming_scan.py::test_scan_model_streaming_symlink_outside_directory_matches_normal_scan tests/test_streaming_scan.py::test_scan_model_streaming_symlink_outside_directory_without_safe_files_returns_security_exit_code tests/test_streaming_scan.py::test_scan_model_streaming_hf_cache_symlink_allowed tests/test_streaming_scan.py::test_scan_model_streaming_hf_home_cache_symlink_allowed tests/test_streaming_scan.py::test_scan_model_streaming_symlink_reports_source_path_consistently tests/test_streaming_scan.py::test_scan_model_streaming_hf_cache_symlink_reports_snapshot_path tests/test_streaming_scan.py::test_scan_model_streaming_oversized_renamed_safetensors_fails_before_hashing tests/utils/file/test_advanced_file_handler.py::TestShardedModelDetector::test_validated_shard_target_mapping_rejects_alias_swap tests/utils/file/test_advanced_file_handler.py::TestShardedModelDetector::test_shard_target_swap_after_detection_fails_closed tests/utils/file/test_advanced_file_handler.py::TestShardedModelDetector::test_shard_target_swap_during_scan_discards_clean_result tests/utils/file/test_advanced_file_handler.py::TestShardedModelDetector::test_shard_target_swap_during_scan_preserves_security_findings tests/utils/file/test_advanced_file_handler.py::TestShardedModelDetector::test_sharded_model_rejects_config_symlink_swap_without_nofollow tests/utils/file/test_advanced_file_handler.py::TestAdvancedFileHandler::test_duplicate_hardlinked_shard_family_members_are_not_cacheable
# 16 passed
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -q tests/test_streaming_scan.py tests/scanners/test_onnx_scanner.py tests/utils/sources/test_huggingface.py
# 604 passed, 1 skipped, 1 warning
uv run ruff check modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
# All checks passed
uv run ruff format --check modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
# 419 files already formatted
uv run mypy modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
# Success: no issues found in 474 source files
git diff --check
# clean

@codex review

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 7975772a51

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".

mldangelo-oai · 2026-06-11T03:21:50Z

CI diagnosis for current head 7975772a: Quick Feedback (3.12), Python 3.13, and Windows all fail the same deterministic regression: tests/test_core.py::test_directory_scan_does_not_reresolve_trusted_hf_alias, where alias_resolve_calls is 2 but expected 1. Please include this exact cross-platform failure in the current follow-up fix and rerun the focused test before pushing.

mldangelo-oai · 2026-06-11T04:06:23Z

@codex review

Fresh exact-head review requested for commit 4ed5fd23b8e9f5672cf957a4529c6e183cc16166.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4ed5fd23b8

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".

mldangelo-oai · 2026-06-11T04:44:28Z

@codex review

Fresh exact-head review requested for commit da12fe667b743bad11e09caf5976ad7ad1291b84.

chatgpt-codex-connector · 2026-06-11T04:48:06Z

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".

mldangelo-oai · 2026-06-11T05:14:09Z

@codex review

CI is green on head da12fe667b743bad11e09caf5976ad7ad1291b84; requesting fresh exact-head review for code-owner approval readiness.

chatgpt-codex-connector · 2026-06-11T05:17:25Z

Codex Review: Didn't find any major issues. Keep it up!

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".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8411e1104f

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".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 500064663d

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".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: efdd9159d7

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".

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9a8f5c676d

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".

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/core.py # tests/test_streaming_scan.py

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f81171fb9b

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".

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py # tests/utils/sources/test_huggingface.py

…t08-onnx-external-data-streaming-20260610

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py

…t08-onnx-external-data-streaming-20260610

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/core.py # tests/test_streaming_scan.py

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f9ff39421e

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".

…t08-onnx-external-data-streaming-20260610

…t08-onnx-external-data-streaming-20260610 # Conflicts: # tests/test_streaming_scan.py

fix: preserve ONNX external data in HF streaming

b7e346a

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/utils/sources/huggingface.py Outdated

Comment thread modelaudit/utils/sources/huggingface.py Outdated

Comment thread modelaudit/utils/sources/huggingface.py Outdated

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

442db47

…t08-onnx-external-data-streaming-20260610

fix: address hf onnx streaming review feedback

dc4542b

mldangelo-oai requested a review from mldangelo June 11, 2026 01:57

mldangelo-oai enabled auto-merge (squash) June 11, 2026 01:57

mldangelo-oai disabled auto-merge June 11, 2026 02:26

fix: resolve ONNX external data from HF cache aliases

7975772

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/utils/sources/huggingface.py Outdated

Comment thread modelaudit/core.py Outdated

Comment thread modelaudit/scanners/onnx_scanner.py

fix: address onnx external data review edge cases

4ed5fd2

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/onnx_scanner.py

Comment thread modelaudit/utils/sources/huggingface.py Outdated

fix: tighten onnx external data edge cases

da12fe6

mldangelo-oai enabled auto-merge (squash) June 11, 2026 05:24

fix: preserve content-routed HF ONNX aliases

8411e11

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/utils/sources/huggingface.py

fix: anchor HF cache snapshot aliases

4971f5a

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/utils/sources/huggingface.py Outdated

Comment thread modelaudit/scanners/onnx_scanner.py Outdated

Comment thread modelaudit/core.py Outdated

fix: account for trusted ONNX sidecars

efdd915

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/core.py Outdated

Comment thread modelaudit/core.py

Comment thread modelaudit/core.py Outdated

fix: cover suffixed ONNX sidecar streaming

9a8f5c6

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/utils/sources/huggingface.py

mldangelo-oai added 3 commits June 11, 2026 20:51

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

d3277f6

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/core.py # tests/test_streaming_scan.py

fix: enforce ONNX sidecar stream limits

e0134da

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

f81171f

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/onnx_scanner.py

mldangelo-oai added 14 commits June 11, 2026 22:57

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

9add036

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py # tests/utils/sources/test_huggingface.py

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

1834c2f

…t08-onnx-external-data-streaming-20260610

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

5956176

…t08-onnx-external-data-streaming-20260610

fix: count shared ONNX sidecar bytes once

75e9e8c

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

350a489

…t08-onnx-external-data-streaming-20260610

fix: dedupe HF ONNX sidecar aliases

208ba51

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

dc6dedd

…t08-onnx-external-data-streaming-20260610

fix: scan scannable ONNX sidecars

6831910

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

edb1288

…t08-onnx-external-data-streaming-20260610

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

441d3f1

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

4c73fb1

…t08-onnx-external-data-streaming-20260610

test: adapt HF streaming selection assertions

31e6dea

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

9341a72

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/core.py # tests/test_streaming_scan.py

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

f9ff394

…t08-onnx-external-data-streaming-20260610 # Conflicts: # modelaudit/utils/sources/huggingface.py

chatgpt-codex-connector Bot reviewed Jun 12, 2026

View reviewed changes

Comment thread modelaudit/core.py

Comment thread modelaudit/core.py

mldangelo-oai added 4 commits June 12, 2026 07:48

fix: hash onnx external data in directory scans

807fd3d

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

1634970

…t08-onnx-external-data-streaming-20260610

test: adapt HF FP selector assertion

4e64939

Merge remote-tracking branch 'origin/main' into mdangelo/codex/hf-fp-…

73bdcbf

…t08-onnx-external-data-streaming-20260610 # Conflicts: # tests/test_streaming_scan.py

mldangelo-oai merged commit 67fd0d4 into main Jun 12, 2026
29 checks passed

mldangelo-oai deleted the mdangelo/codex/hf-fp-t08-onnx-external-data-streaming-20260610 branch June 12, 2026 09:36

Conversation

mldangelo-oai commented Jun 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Root Cause

Security Tradeoffs

Pinned BGE-M3 QA

Validation

Follow-up Validation (4ed5fd2)

Follow-up Validation (da12fe6)

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

github-actions Bot commented Jun 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Performance Benchmarks

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

chatgpt-codex-connector Bot commented Jun 11, 2026

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

chatgpt-codex-connector Bot commented Jun 11, 2026

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

mldangelo-oai commented Jun 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

chatgpt-codex-connector Bot commented Jun 11, 2026

Uh oh!

mldangelo-oai commented Jun 11, 2026

Uh oh!

chatgpt-codex-connector Bot commented Jun 11, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

mldangelo-oai commented Jun 11, 2026 •

edited

Loading

Follow-up Validation (`4ed5fd2`)

Follow-up Validation (`da12fe6`)

github-actions Bot commented Jun 11, 2026 •

edited

Loading

mldangelo-oai commented Jun 11, 2026 •

edited

Loading