fix(text): deduplicate model-card indicators by mldangelo-oai · Pull Request #1631 · promptfoo/modelaudit

mldangelo-oai · 2026-06-10T23:56:26Z

Summary

correlate same-span model-card URL, domain, cloud-storage, and git-clone evidence into one normalized fingerprinted finding
keep distinct executable indicators, nested/encoded URLs, suspicious ports, credentials, locations, code examples, and malicious controls separate
keep passive Markdown/table documentation links informational and feed evidence fingerprints into SARIF deduplication

Validation

uv run ruff format modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
uv run ruff check --fix modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/
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/
git diff --check
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/integrations/test_sarif_formatter.py -q (510 passed)
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/detectors/test_network_comm_detector.py tests/integrations/test_sarif_formatter.py tests/test_network_comm_integration.py -q (1043 passed, 9 skipped)
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1 (17379 passed, 1292 skipped, 39 warnings)
pinned Hugging Face cards: sentence-transformers/all-MiniLM-L6-v2@1110a243fdf4706b3f48f1d95db1a4f5529b4d41, openai/clip-vit-base-patch32@3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268, microsoft/resnet-50@34c2154c194f829b11125337b98c8f5f9965ff19; doc-only cards exit 0, CLIP code example remains actionable, and paired malicious controls retain one grouped git-clone finding plus separate requests.get control

github-actions · 2026-06-10T23:58:26Z

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: 1.459s -> 1.459s (+0.1%).

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	104.80ms	111.36ms	+6.3%	stable
`clean-training-checkpoint`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_clean_training_checkpoint`	`safe_large`	278.2 KiB	1	115.49ms	111.78ms	-3.2%	stable
`chunked-upload-stream`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_chunked_upload_stream`	`chunked_stream`	278.2 KiB	1	118.95ms	115.29ms	-3.1%	stable
`direct-malicious-upload`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_direct_malicious_upload`	`malicious_reduce`	52 B	1	463.2us	473.6us	+2.3%	stable
`nested-payload-review`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_base64]`	`nested_base64`	98 B	1	520.0us	511.5us	-1.6%	stable
`padded-multi-stream-upload`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_padded_multi_stream_upload`	`multi_stream_padded`	4.1 KiB	1	587.1us	578.3us	-1.5%	stable
`nested-payload-review`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_raw]`	`nested_raw`	78 B	1	518.7us	512.3us	-1.2%	stable
`suspicious-pickle-intake`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_suspicious_pickle_intake`	`suspicious-intake`	183.8 KiB	4	144.44ms	145.47ms	+0.7%	stable
`nested-payload-review`	`tests/benchmarks/test_picklescan_benchmarks.py::test_picklescan_nested_payload_review[nested_hex]`	`nested_hex`	130 B	1	548.0us	550.3us	+0.4%	stable
`single-checkpoint-preflight`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_single_checkpoint_before_load`	`single_checkpoint.pkl`	183.0 KiB	1	74.19ms	73.92ms	-0.4%	stable
`mixed-model-repository`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_release_candidate_repository`	`release-candidate`	547.3 KiB	32	488.33ms	489.24ms	+0.2%	stable
`duplicate-heavy-registry`	`tests/benchmarks/test_scan_benchmarks.py::test_scan_duplicate_registry_snapshot`	`registry-snapshot`	915.2 KiB	13	409.71ms	409.59ms	-0.0%	stable

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4c6f571434

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 · 2026-06-11T00:50:27Z

@codex review

chatgpt-codex-connector · 2026-06-11T00:53:47Z

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 · 2026-06-11T06:23:41Z

@codex review

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 444d0c3e73

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 · 2026-06-11T06:55:24Z

@codex review

Updated head: e208fb4.

Addressed the three prior P2 findings:

SARIF primary fingerprints now combine evidence fingerprint, redacted artifact path, and severity, while preserving the scanner evidence fingerprint in properties.
Model-card evidence line/column lookup is precomputed for finding positions so dedup no longer counts newlines from the start for every finding.
Passive Markdown-link suppression now excludes code contexts such as endpoint assignments; added regression for endpoint = "[download](https://evil.example/payload.sh)".

Validation on this head:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/integrations/test_sarif_formatter.py -q -> 512 passed
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/detectors/test_network_comm_detector.py tests/integrations/test_sarif_formatter.py tests/test_network_comm_integration.py -q -> 1045 passed, 9 skipped
uv run ruff check --fix ..., uv run ruff check ..., uv run ruff format --check ..., uv run mypy ..., git diff --check all clean
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1 -> 17380 passed, 1292 skipped
Pinned HF README cards at exact revisions: MiniLM and ResNet doc-only cards exit 0, CLIP code example remains actionable, injected git-clone/requests controls retained separately through JSON and SARIF.

mldangelo-oai · 2026-06-11T06:57:33Z

Pinned real-model-card QA from Hugging Face rank 255 exposes an additional benign coverage failure on current main.

Model/revision: openai/whisper-large-v2@ae4642769ce2ad8fc292556ccea8e901f1530655
ModelAudit: 8d6c4864fe2ea833ceaef1b9803d225afb1e8d69
Result: 40 informational S309 README URL/domain findings, then Text Content Security Coverage fails because endpoint redaction classification exceeds max_classifications=32.
Outcome: analysis_incomplete=true, text_content_security_finding_limit; the overall model scan exits 2.

The model card is benign documentation. Please use this immutable revision as real-world QA for dedup/calibration and prove the exact PR head does not exhaust the classification budget merely because the README contains many ordinary documentation links. The fix must preserve actionable code-context/network findings and bounded work rather than simply raising the cap.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e208fb43b5

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 · 2026-06-11T07:27:04Z

@codex review

Updated head: bfad694.

Addressed the new Whisper real-card QA comment:

Documentation-only endpoint-redaction classification exhaustion is now reported as an informational reporting limit when the inspected README contains only passive documentation network-token lines.
Actionable code/network token lines still keep the detector limit fail-closed.
BibTeX bibliography fields such as url = {https://arxiv.org/...} and citation copyright domains are treated as passive documentation, while normal code assignments remain actionable.

Validation on this head:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py -q -> 446 passed
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/integrations/test_sarif_formatter.py -q -> 515 passed
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/detectors/test_network_comm_detector.py tests/integrations/test_sarif_formatter.py tests/test_network_comm_integration.py -q -> 1048 passed, 9 skipped
uv run ruff check ..., uv run ruff format --check ..., uv run mypy ..., git diff --check all clean
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1 -> 17383 passed, 1292 skipped
Pinned HF README cards at exact revisions: MiniLM and ResNet doc-only cards exit 0; CLIP real code example remains actionable; Whisper doc-only README now exits 0; injected git-clone/requests controls are retained through JSON/SARIF on all four cards, with Whisper+controls failing closed due actionable code after the redaction budget.

chatgpt-codex-connector · 2026-06-11T07:30:43Z

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 · 2026-06-11T07:43:28Z

@codex review

Updated head: eb2acc8734fb17b5797c08e6a80ce9cd43f7850d.

Addressed the new P2 return-context finding:

Markdown links inside returned/yielded strings are now treated as code context instead of passive documentation.
Added regression for return "[download](https://evil.example/payload.sh)"; it stays actionable and exits 1.

Validation on this head:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py -q -> 447 passed
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/detectors/test_network_comm_detector.py tests/integrations/test_sarif_formatter.py tests/test_network_comm_integration.py -q -> 1049 passed, 9 skipped
uv run ruff check ..., uv run ruff format --check ..., uv run mypy ..., git diff --check all clean
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1 -> 17384 passed, 1292 skipped
Pinned HF README matrix still passes: MiniLM/ResNet/Whisper doc-only exit 0; CLIP code example remains actionable; injected git-clone/requests controls retained through JSON/SARIF on all four cards.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: eb2acc8753

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 · 2026-06-11T08:10:53Z

@codex review

Updated head: 541911d.

Addressed the three fresh review findings:

Parenthesized returned/yielded Markdown-link strings now stay in code context, e.g. return ("[download](https://evil.example/payload.sh)") remains actionable.
BibTeX passive-field handling now only suppresses recognized passive bibliography fields, so an unclosed @misc{... no longer downgrades endpoint = "https://evil.example/payload.sh".
Model-card evidence columns now count decoded text characters rather than UTF-8 bytes before the evidence span.

Validation on this head:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py -q -k 'unicode_characters or parenthesized_markdown_link_return or unclosed_bibliography' -> 3 passed, 447 deselected
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/detectors/test_network_comm_detector.py tests/integrations/test_sarif_formatter.py tests/test_network_comm_integration.py -q -> 1052 passed, 9 skipped
uv run ruff format modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/ -> 2 files reformatted before commit
uv run ruff check --fix modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/ -> all checks passed
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 in 474 source files
git diff --check -> clean
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1 -> 17387 passed, 1292 skipped
Pinned real-card matrix (--scanners text --no-cache, JSON and SARIF for benign/control variants): MiniLM benign 0/control 2, CLIP benign 1/control 1, ResNet benign 0/control 1, Whisper benign 0/control 2; all controlled variants retained evil.example evidence in JSON and SARIF.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 541911d973

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 · 2026-06-11T08:35:33Z

@codex review

Updated head: 8dea686.

Addressed the two fresh P1 findings:

Quoted url = "https://..." assignments are no longer treated as passive BibTeX fields under an unclosed @misc{...; brace-style bibliography URL fields remain informational.
URLs in indented multiline return/yield values, including returned Markdown links split after return (, are now treated as code context before passive Markdown-link downgrading and before documentation-only reporting-limit handling.

Validation on this head:

PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py -q -k 'multiline_markdown_link_return or quoted_url_code or unclosed_bibliography or bibliography_url_field' -> 4 passed, 448 deselected
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest tests/scanners/test_text_scanner.py tests/detectors/test_network_comm_detector.py tests/integrations/test_sarif_formatter.py tests/test_network_comm_integration.py -q -> 1054 passed, 9 skipped
uv run ruff format modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/ -> 419 files left unchanged
uv run ruff check --fix modelaudit/ packages/modelaudit-picklescan/src packages/modelaudit-picklescan/tests tests/ -> all checks passed
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 in 474 source files
git diff --check -> clean
PROMPTFOO_DISABLE_TELEMETRY=1 uv run pytest -n auto -m "not slow and not integration" --maxfail=1 -> 17389 passed, 1292 skipped
Pinned real-card matrix (--scanners text --no-cache, JSON and SARIF for benign/control variants, with quoted url = ... and multiline return controls): MiniLM benign 0/control 2, CLIP benign 1/control 1, ResNet benign 0/control 1, Whisper benign 0/control 2; all controlled variants retained evil.example evidence in JSON and SARIF.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8dea686941

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 · 2026-06-11T10:48:04Z

Codex Review: Didn't find any major issues. Nice work!

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 · 2026-06-11T11:05:27Z

@codex review

Head: cd43020

Addressed the two code-quality redundant-comparison threads:

Direct parent-line branch now returns True after the existing line_indent < value_indent guard.
List-object parent lookup now returns only the meaningful parent-line pattern check.

Validation on this head:

focused endpoint-redaction/nested-config subset: 6 passed
affected text/network/SARIF suite: 1061 passed, 9 skipped
ruff check and ruff format --check
mypy over modelaudit, picklescan src/tests, tests
git diff --check
full fast lane: 17396 passed, 1292 skipped
pinned real-model-card JSON/SARIF matrix with plain nested config, YAML list-item config, YAML list-object config, encoded nested URL, returned/wrapped markdown links, redacted credential, and controlled malicious URLs

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: cd430203b4

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 · 2026-06-11T12:03:15Z

@codex review

Head: e3b265e

Validation run on this head:

Focused text-scanner regressions: 13 passed
Affected scanner/network/SARIF suite: 1065 passed, 9 skipped
Ruff format, Ruff check, mypy, and git diff --check: clean
Full fast lane: 17400 passed, 1292 skipped
Pinned real-model-card JSON/SARIF matrix: passed (MiniLM 0/1, CLIP 1/1, ResNet 0/1, Whisper 0/2 expected exits)

chatgpt-codex-connector · 2026-06-11T12:07:00Z

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

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 · 2026-06-11T12:24:30Z

@codex review

Head: 254b6a5e8619f0f057f5e9e7aca1cf8a80ff7c9e

Addressed the code-quality cleanup on the previous head by removing the now-dead absolute_position assignment in the domain-token branch.

Validation run on this head:

Focused text-scanner regressions: 13 passed
Affected scanner/network/SARIF suite: 1065 passed, 9 skipped
Ruff format, Ruff check, mypy, and git diff --check: clean
Pinned real-model-card JSON/SARIF matrix: passed (MiniLM 0/1, CLIP 1/1, ResNet 0/1, Whisper 0/2 expected exits)
Full fast lane: 17400 passed, 1292 skipped

chatgpt-codex-connector · 2026-06-11T12:28:45Z

Codex Review: Didn't find any major issues. Swish!

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: 6ed93d9505

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".

…t04-model-card-finding-dedup-20260610

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 7c28e9a2b3

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".

…t04-model-card-finding-dedup-20260610

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 771d073b17

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".

…dangelo/codex/hf-fp-t04-model-card-finding-dedup-20260610

fix(text): deduplicate model-card indicators

4c6f571

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

fix(text): preserve grouped indicator severity

dbf58c3

mldangelo-oai requested a review from mldangelo June 11, 2026 01:24

mldangelo-oai enabled auto-merge (squash) June 11, 2026 01:24

mldangelo-oai added 2 commits June 11, 2026 05:51

Merge remote-tracking branch 'origin/main' into HEAD

4025502

fix(text): preserve distinct model-card evidence

444d0c3

mldangelo-oai disabled auto-merge June 11, 2026 06:23

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/integrations/sarif_formatter.py

Comment thread modelaudit/scanners/text_scanner.py

Comment thread modelaudit/scanners/text_scanner.py

fix(text): address model-card evidence review

e208fb4

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py

fix(text): treat passive bibliography links as documentation

bfad694

fix(text): keep returned markdown links actionable

eb2acc8

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

Comment thread modelaudit/scanners/text_scanner.py Outdated

Comment thread modelaudit/scanners/text_scanner.py Outdated

fix(text): address model-card context review

541911d

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

Comment thread modelaudit/scanners/text_scanner.py

fix(text): preserve multiline model-card code context

8dea686

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

Comment thread modelaudit/scanners/text_scanner.py

Comment thread modelaudit/scanners/text_scanner.py Outdated

fix(text): fail closed on wrapped model-card calls

ee9c135

github-code-quality Bot found potential problems Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Fixed

Comment thread modelaudit/scanners/text_scanner.py Fixed

fix(text): simplify nested config context checks

cd43020

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

Comment thread modelaudit/scanners/text_scanner.py Outdated

Comment thread modelaudit/scanners/text_scanner.py Outdated

fix(text): handle bibliography endpoint edge cases

e3b265e

github-code-quality Bot found potential problems Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Fixed

chore(text): remove redundant position assignment

254b6a5

mldangelo-oai enabled auto-merge (squash) June 11, 2026 12:56

fix(text): preserve markdown links in code literals

6ed93d9

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

mldangelo-oai added 2 commits June 11, 2026 16:56

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

15e3db1

…t04-model-card-finding-dedup-20260610

fix(text): handle zero-indent table contexts

7c28e9a

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

mldangelo-oai added 3 commits June 11, 2026 18:37

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

75c8e05

…t04-model-card-finding-dedup-20260610

fix(text): handle CRLF multiline literals

80ab47a

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

771d073

…t04-model-card-finding-dedup-20260610

chatgpt-codex-connector Bot reviewed Jun 11, 2026

View reviewed changes

Comment thread modelaudit/scanners/text_scanner.py Outdated

mldangelo-oai added 5 commits June 12, 2026 01:05

Merge branch 'main' of https://github.com/promptfoo/modelaudit into m…

0cbd1b2

…dangelo/codex/hf-fp-t04-model-card-finding-dedup-20260610

fix(text): preserve structured markdown contexts

cb81d01

Merge branch 'main' of https://github.com/promptfoo/modelaudit into m…

205c7fa

…dangelo/codex/hf-fp-t04-model-card-finding-dedup-20260610

fix(text): preserve zero-indent literal links

2cf54d4

Merge branch 'main' of https://github.com/promptfoo/modelaudit into m…

fbb7855

…dangelo/codex/hf-fp-t04-model-card-finding-dedup-20260610

mldangelo-oai merged commit 5ff4247 into main Jun 12, 2026
29 checks passed

mldangelo-oai deleted the mdangelo/codex/hf-fp-t04-model-card-finding-dedup-20260610 branch June 12, 2026 03:35

github-actions Bot mentioned this pull request Jun 24, 2026

chore: release main #1713

Merged

Conversation

mldangelo-oai commented Jun 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Validation

Uh oh!

github-actions Bot commented Jun 10, 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!

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 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

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

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 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 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 left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

chatgpt-codex-connector Bot commented Jun 11, 2026

Uh oh!

Uh oh!

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 10, 2026 •

edited

Loading

github-actions Bot commented Jun 10, 2026 •

edited

Loading