fix(security): raise pypdf floor to 6.16.1 for CVE-2026-84309/84310/84311

CHANGELOG.md

@@ -35,6 +35,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - API 에러 응답 생성 시 내부 예외 체인을 억제하여 의존성 오류나 내부 경로가 노출될 가능성을 줄임
 - API 응답 미들웨어에 `Cache-Control: no-store, max-age=0` 헤더를 추가하여 민감한 파싱 데이터의 브라우저 및 중간 캐싱을 방지
 - `uv.lock`의 의존성을 재잠금하여 실제 `pip-audit`/`trivy-fs` CVE를 제거: 런타임 경로의 `pillow` 12.2.0→12.3.0 (PYSEC-2026-3451/3452/3453/3454/3493/3494/3495/3496, 이미지 파서 취약점 8건), `pypdf>=6.15.0,<7.0` (lock 6.15.0; CVE-2026-59935/59936/59937/59938/71852/71870, PDF 파싱 경로), `click` 8.3.2→8.4.2 (PYSEC-2026-2132) — 모두 스캔 PDF/이미지 파싱 런타임에 직접 관련되며 선언 범위와 lock을 함께 고정함. 빌드 도구 `setuptools` 81.0.0→83.0.0 (CVE-2026-59890). 문서 툴체인의 `pymdown-extensions` 10.21.3→11.0.1 (CVE-2026-61632, MEDIUM)은 `mkdocs-material` 9.6.x의 `pymdown-extensions~=10.2`(`<11`) 상한 때문에 막혀 있었으므로, docs extra 핀을 `mkdocs-material>=9.7,<9.8`로 올려(9.7.x는 상한을 `>=10.2`로 완화) 해소함. `uv run mkdocs build --strict` 통과 확인. 조치 후 전체 잠금(런타임+extras) `pip-audit`: 취약점 0건.
+- `pypdf`를 `>=6.16.1,<7.0`(lock 6.16.2)로 재잠금하여 신규 `trivy-fs` MEDIUM 3건을 제거: CVE-2026-84309(순환 `/Next` 아웃라인 트리로 `TreeObject.insert_child`가 무한 루프에 빠짐), CVE-2026-84310(`_get_outline`의 항목 수·중첩 깊이 제한 부재로 장시간·과다 메모리 소비), CVE-2026-84311(재사용 XObject 폼 그래프에 방문 여부 추적이 없어 텍스트 추출 시 경로가 지수적으로 증가) — 모두 신뢰할 수 없는 PDF 업로드를 처리하는 파서 가용성 취약점이며 pypdf 6.16.0/6.16.1에서 각각 수정됨. 조치 후 전체 잠금 `pip-audit`: 취약점 0건.
 
 ### Performance
 - `newsdom_api.dom_builder._html_safe_text` 함수에 early return과 타입 체크를 도입하여 불필요한 `str()` 캐스팅을 제거함으로써 처리 속도를 개선했습니다.

docs/doctoring/dependency-security-baseline.md

@@ -14,12 +14,12 @@ The adopted floors are:
 
 - `setuptools>=83` for the build backend;
 - `Pillow>=12.3,<13.0` for image parsing on the untrusted document-ingestion path;
-- `pypdf>=6.15.0,<7.0` for PDF parsing;
+- `pypdf>=6.16.1,<7.0` for PDF parsing;
 - `mkdocs-material>=9.7,<9.8`, allowing `pymdown-extensions>=11` while the MkDocs
   core remains on the supported 1.x line.
 
 The generated lock additionally resolves Click 8.4.2, setuptools 83.0.0,
-Pillow 12.3.0, pypdf 6.15.0, mkdocs-material 9.7.7, and
+Pillow 12.3.0, pypdf 6.16.2, mkdocs-material 9.7.7, and
 pymdown-extensions 11.0.1. Direct floors prevent a later lock refresh from
 silently selecting known-vulnerable ranges again.
 
@@ -36,6 +36,20 @@ successfully without suppressing either finding. The shared direct floor and loc
 therefore move together to 6.15.0 rather than hiding the findings in
 `.trivyignore`.
 
+On September 3, 2026, an unrelated PR's `trivy-fs` gate reported three further
+MEDIUM findings against the locked 6.15.0 artifact: CVE-2026-84309 (an attacker
+-crafted PDF whose cyclic `/Next` outline-tree structure drives
+`TreeObject.insert_child` into an infinite loop on a writing code path),
+CVE-2026-84310 (unbounded entry-count and nesting-depth traversal in
+`_get_outline` allowing long runtimes and large memory use), and CVE-2026-84311
+(a reused-XObject form graph without a memoized/visited guard in
+`PageObject._extract_text`/`extract_xform_text`, producing exponentially many
+traversal paths). All three are availability risks directly on NewsDOM's
+untrusted PDF-ingestion path and are fixed upstream in pypdf 6.16.0/6.16.1
+respectively. The direct floor and lock move together to 6.16.1 (lock resolves
+6.16.2, the latest compatible release at the time of this record) rather than
+suppressing the findings.
+
 CVE-2026-59890 affects setuptools versions before 83.0.0. On
 normalization-preserving macOS filesystems, specially named files could bypass
 `MANIFEST.in` exclusion matching and enter a source distribution. Although this
@@ -46,7 +60,7 @@ Pillow 12.3.0 and pypdf release artifacts are distributed through PyPI with
 published cryptographic file digests. Those artifacts and digests provide
 provenance inputs; they do not by themselves establish that a package is safe.
 Repository scans, hash-locked resolution, current-head tests, and independent
-review remain mandatory. PyPI's official JSON metadata confirms the 6.15.0
+review remain mandatory. PyPI's official JSON metadata confirms the 6.16.2
 release and the artifact hashes recorded in this repository's generated lock.
 
 ## Secure-development and provenance controls
@@ -129,6 +143,15 @@ Open Source Vulnerabilities. (2026c). *CVE-2026-71852*. Retrieved August 9,
 Open Source Vulnerabilities. (2026d). *CVE-2026-71870*. Retrieved August 9,
     2026, from https://osv.dev/vulnerability/CVE-2026-71870
 
+Open Source Vulnerabilities. (2026e). *CVE-2026-84309*. Retrieved September 3,
+    2026, from https://osv.dev/vulnerability/CVE-2026-84309
+
+Open Source Vulnerabilities. (2026f). *CVE-2026-84310*. Retrieved September 3,
+    2026, from https://osv.dev/vulnerability/CVE-2026-84310
+
+Open Source Vulnerabilities. (2026g). *CVE-2026-84311*. Retrieved September 3,
+    2026, from https://osv.dev/vulnerability/CVE-2026-84311
+
 Python Packaging Authority. (2026a). *Digital attestations*. PyPI Docs.
     Retrieved August 4, 2026, from https://docs.pypi.org/attestations/
 
@@ -140,3 +163,6 @@ Python Packaging Authority. (2026c). *pypdf 6.15.0*. Python Package Index.
 
 Python Packaging Authority. (2026d). *setuptools 83.0.0*. Python Package Index.
     Retrieved August 4, 2026, from https://pypi.org/project/setuptools/83.0.0/
+
+Python Packaging Authority. (2026e). *pypdf 6.16.2*. Python Package Index.
+    Retrieved September 3, 2026, from https://pypi.org/project/pypdf/6.16.2/

pyproject.toml

@@ -17,7 +17,7 @@ dependencies = [
   "python-multipart>=0.0.31,<1.0",
   "reportlab>=4.2,<6.0",
   "Pillow>=12.3,<13.0",
-  "pypdf>=6.15.0,<7.0",
+  "pypdf>=6.16.1,<7.0",
 ]
 
 [project.optional-dependencies]

tests/test_project_metadata.py

@@ -96,7 +96,7 @@ def test_security_dependency_floors_exclude_known_vulnerable_ranges():
     dependencies_section = _dependencies_section(text)
 
     assert '"Pillow>=12.3,<13.0"' in dependencies_section
-    assert '"pypdf>=6.15.0,<7.0"' in dependencies_section
+    assert '"pypdf>=6.16.1,<7.0"' in dependencies_section
     assert 'requires = ["setuptools>=83", "wheel"]' in text
 
 
@@ -202,4 +202,4 @@ def test_uv_lock_does_not_track_external_mineru_pipeline_runtime_stack():
 
 
 def test_uv_lock_pins_pypdf_at_patched_release():
-    assert _locked_package_version("pypdf") >= (6, 15, 0)
+    assert _locked_package_version("pypdf") >= (6, 16, 1)

tests/test_pypdf_security_floor.py

@@ -6,9 +6,13 @@
 import yaml
 
 
-_REQUIRED_PYPDF_VERSION = (6, 15, 0)
-_CURRENT_PYPDF_CVES = ("CVE-2026-71852", "CVE-2026-71870")
-_LOCKED_PYPDF_REQUIREMENT = '{ name = "pypdf", specifier = ">=6.15.0,<7.0" },'
+_REQUIRED_PYPDF_VERSION = (6, 16, 1)
+_CURRENT_PYPDF_CVES = ("CVE-2026-84309", "CVE-2026-84310", "CVE-2026-84311")
+# Every literal below derives from the tuple above, so a future floor bump has
+# exactly one place to change and the documentation checks cannot keep
+# matching a superseded floor by accident.
+_PYPDF_SPECIFIER = ">=" + ".".join(str(part) for part in _REQUIRED_PYPDF_VERSION) + ",<7.0"
+_LOCKED_PYPDF_REQUIREMENT = f'{{ name = "pypdf", specifier = "{_PYPDF_SPECIFIER}" }},'
 
 
 def _locked_pypdf_version() -> tuple[int, ...]:
@@ -24,10 +28,10 @@ def _locked_pypdf_version() -> tuple[int, ...]:
 
 
 def test_project_declares_current_pypdf_security_floor() -> None:
-    """Prevent future lock refreshes from selecting the vulnerable 6.14.x line."""
+    """Prevent lock refreshes from selecting pypdf below the current floor."""
 
     project_text = Path("pyproject.toml").read_text(encoding="utf-8")
-    assert '"pypdf>=6.15.0,<7.0"' in project_text
+    assert f'"pypdf{_PYPDF_SPECIFIER}"' in project_text
 
 
 def test_lock_uses_current_pypdf_security_release() -> None:
@@ -61,7 +65,25 @@ def test_current_pypdf_advisories_and_floor_are_documented() -> None:
 
     for cve_id in _CURRENT_PYPDF_CVES:
         assert f"https://osv.dev/vulnerability/{cve_id}" in baseline
-    assert "`pypdf>=6.15.0,<7.0`" in changelog
+
+    # The CHANGELOG entry writes the package and its specifier as separate
+    # inline-code spans; require both, plus the release the lock resolved,
+    # on one entry line rather than a single fused code span.
+    locked_release = ".".join(str(part) for part in _locked_pypdf_version())
+    changelog_entry = re.compile(
+        r"`pypdf`[^\n]*`" + re.escape(_PYPDF_SPECIFIER) + r"`[^\n]*" + re.escape(locked_release)
+    )
+    assert changelog_entry.search(changelog), (
+        f"CHANGELOG.md lacks an entry naming `pypdf`, `{_PYPDF_SPECIFIER}` and lock {locked_release}"
+    )
+
+    # Markdown wraps the provenance sentence across lines; compare on
+    # whitespace-normalized text, and tie the release to the lock file.
+    provenance = (
+        f"PyPI's official JSON metadata confirms the {locked_release} release and the "
+        "artifact hashes recorded in this repository's generated lock."
+    )
+    assert provenance in re.sub(r"\s+", " ", baseline)
 
 
 def test_trivy_registry_exception_is_scoped_to_the_example_manifest() -> None: