Hugging Face's logo

Datasets:
DimensionV
/
KR-Mobile-Smishing-SIEM-Dataset-2026

Dataset card Files
xet
 Community
Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot extract the features (columns) for the split 'train' of the config 'default' of the dataset.
Error code:   FeaturesError
Exception:    ArrowInvalid
Message:      Schema at index 1 was different: 
incident_id: string
case_type: string
theme: string
difficulty_tier: string
timeline_profile: string
start_time: timestamp[s]
end_time: timestamp[s]
platform_coverage: list<item: string>
minimum_log_sources_required: list<item: string>
full_log_sources_used: list<item: string>
noise_model_applied: struct<profile: string, background_event_ratio: double>
message_samples_used: list<item: struct<target_device: string, text: string, sender_type: string, delivery_channel: string>>
page_construction_detail_safe: struct<no_brand_copy: bool, stepper_ui_flow: list<item: string>, visual_style_notes: list<item: string>, form_risk_class: string, fields_requested: list<item: string>, detection_cues_for_human_analyst: list<item: string>>
persona: struct<user_id: string, job_title: string, risk_profile: string, devices: list<item: struct<device_id: string, platform: string, os_version: string>>>
behavioral_model_for_this_incident: struct<scenario: string, why_this_behavior: string, time_gap_rationale: string>
expected_rules: struct<must_fire: list<item: string>, should_fire: list<item: string>, optional: list<item: null>, expected_fire_path: struct<SMISH_REDIRECT_CHAIN: struct<platform: string, trigger_events: list<item: string>>, SENSITIVE_FORM_SEQUENCE: struct<platform: string, trigger_events: list<item: string>>, POST_LURE_FINANCIAL_RISK_SIGNAL: struct<platform: string, trigger_events: list<item: string>>>>
event_stream: list<item: struct<_event_index: string, _phase: string, event_id: string, time: timestamp[s], log_source: string, event_type: string, severity: string, entity: struct<user_id: string, device_id: string, platform: string, os_version: string, mdm_enrolled: bool, device_fingerprint_status: string>, network: struct<carrier: string, carrier_network_type: string, roaming: bool, dst_domain: string, dst_ip: string, ttl_sec: int64, domain_age_days: int64, domain_reputation: double, resolver: string, http: struct<method: string, path: string, status_code: int64>, src_ip: string>, object: struct<sender_type: string, sender_id: string, contains_url: bool, message_features: struct<urgency_language: bool, sender_verification: bool>, qtype: string, traffic_category: string, redirect_hops_total: int64, referrer_chain_domains_masked: list<item: string>, behavior: struct<signal: string, user_interaction: string, dwell_time_ms: int64, form_focus_detected: bool, tap_next: string>, user_initiated: bool, auth: struct<method: string, challenge_id: string, result: string, reason: string, burst_counter_count: int64>, risk_decision: string, result: string, fraud_score: double, flag_reason: string>, correlation: struct<incident_id: string, message_id: string, session_id: string, redirect_chain_id: string, bank_corr_id: string>, label: struct<ground_truth: string, case_type: string, attack_stage: string, mitre_mobile: string, theme: string>, original_source_hint: string, noise_metadata: struct<duplicate_of: string, reason: string>>>
event_stream_timeline_summary: struct<total_events: int64, malicious_events: int64, benign_context_events: int64, noise_events_included: int64, phases: list<item: struct<phase: string, time_range: string, events: int64, desc: string>>, total_elapsed_minutes: double, key_time_gaps_explained: struct<Initial_Response: string, Session_Gap: string, Impact_Delay: string>>
rule_fire_verification: struct<SMISH_REDIRECT_CHAIN: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PHISH_WARNING_BYPASS: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, SENSITIVE_FORM_SEQUENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PLATFORM_DIVERGENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, POST_LURE_FINANCIAL_RISK_SIGNAL: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>>
scaling_template: struct<to_reach_200_events: struct<background_events_to_add: int64, background_composition: struct<messaging_sync: int64, email_sync: int64, push_notification: int64, software_update: int64, cdn_traffic: int64, utility_widgets: int64, mdm_checkin: int64, cloud_backup: int64>>>
vs
incident_id: string
case_type: string
theme: string
difficulty_tier: string
timeline_profile: string
start_time: timestamp[s]
end_time: timestamp[s]
platform_coverage: list<item: string>
minimum_log_sources_required: list<item: string>
full_log_sources_used: list<item: string>
noise_model_applied: struct<profile: string, background_event_ratio: double>
message_samples_used: list<item: struct<target_device: string, text: string, sender_type: string, delivery_channel: string>>
page_construction_detail_safe: struct<no_brand_copy: bool, stepper_ui_flow: list<item: string>, visual_style_notes: list<item: string>, form_risk_class: string, fields_requested: list<item: null>, detection_cues_for_human_analyst: list<item: null>>
persona: struct<user_id: string, job_title: string, risk_profile: string, devices: list<item: struct<device_id: string, platform: string, os_version: string>>>
behavioral_model_for_this_incident: struct<scenario: string, why_this_behavior: string, time_gap_rationale: string>
expected_rules: struct<must_fire: list<item: null>, should_fire: list<item: null>, optional: list<item: null>, expected_fire_path: struct<>>
event_stream: list<item: struct<_event_index: string, _phase: string, event_id: string, time: timestamp[s], log_source: string, event_type: string, severity: string, entity: struct<user_id: string, device_id: string, platform: string, os_version: string, mdm_enrolled: bool, device_fingerprint_status: string, browser: string>, network: struct<carrier: string, carrier_network_type: string, roaming: bool, dst_domain: string, dst_ip: string, ttl_sec: int64, domain_age_days: int64, domain_reputation: double, resolver: string, sni: string, tls_version: string, http: struct<method: string, path: string, status_code: int64, content_type: string, latency_ms: int64>>, object: struct<sender_type: string, sender_id: string, contains_url: bool, message_features: struct<urgency_language: bool, sender_verification: bool>, traffic_category: string, qtype: string, user_initiated: bool, posture: string, secure_dns_or_vpn: bool, os_patch_level: timestamp[s]>, correlation: struct<incident_id: string, message_id: string, session_id: string>, label: struct<ground_truth: string, case_type: string, attack_stage: null, activity_stage: string, theme: string>, original_source_hint: string, noise_metadata: struct<duplicate_of: string, reason: string>>>
event_stream_timeline_summary: struct<total_events: int64, malicious_events: int64, benign_context_events: int64, noise_events_included: int64, phases: list<item: struct<phase: string, time_range: string, events: int64, desc: string>>, total_elapsed_minutes: double, key_time_gaps_explained: struct<sms_to_click: string, page_visit_to_end: string>>
rule_fire_verification: struct<SMISH_REDIRECT_CHAIN: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PHISH_WARNING_BYPASS: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, SENSITIVE_FORM_SEQUENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PLATFORM_DIVERGENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, POST_LURE_FINANCIAL_RISK_SIGNAL: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>>
scaling_template: struct<to_reach_200_events: struct<background_events_to_add: int64, background_composition: struct<messaging_sync: int64, email_sync: int64, push_notification: int64, software_update: int64, cdn_traffic: int64, utility_widgets: int64, mdm_checkin: int64, cloud_backup: int64>>>
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/split/first_rows.py", line 243, in compute_first_rows_from_streaming_response
                  iterable_dataset = iterable_dataset._resolve_features()
                                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 3608, in _resolve_features
                  features = _infer_features_from_batch(self.with_format(None)._head())
                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2368, in _head
                  return next(iter(self.iter(batch_size=n)))
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2573, in iter
                  for key, example in iterator:
                                      ^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2060, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2082, in _iter_arrow
                  yield from self.ex_iterable._iter_arrow()
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 604, in _iter_arrow
                  yield new_key, pa.Table.from_batches(chunks_buffer)
                                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "pyarrow/table.pxi", line 5039, in pyarrow.lib.Table.from_batches
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
              pyarrow.lib.ArrowInvalid: Schema at index 1 was different: 
              incident_id: string
              case_type: string
              theme: string
              difficulty_tier: string
              timeline_profile: string
              start_time: timestamp[s]
              end_time: timestamp[s]
              platform_coverage: list<item: string>
              minimum_log_sources_required: list<item: string>
              full_log_sources_used: list<item: string>
              noise_model_applied: struct<profile: string, background_event_ratio: double>
              message_samples_used: list<item: struct<target_device: string, text: string, sender_type: string, delivery_channel: string>>
              page_construction_detail_safe: struct<no_brand_copy: bool, stepper_ui_flow: list<item: string>, visual_style_notes: list<item: string>, form_risk_class: string, fields_requested: list<item: string>, detection_cues_for_human_analyst: list<item: string>>
              persona: struct<user_id: string, job_title: string, risk_profile: string, devices: list<item: struct<device_id: string, platform: string, os_version: string>>>
              behavioral_model_for_this_incident: struct<scenario: string, why_this_behavior: string, time_gap_rationale: string>
              expected_rules: struct<must_fire: list<item: string>, should_fire: list<item: string>, optional: list<item: null>, expected_fire_path: struct<SMISH_REDIRECT_CHAIN: struct<platform: string, trigger_events: list<item: string>>, SENSITIVE_FORM_SEQUENCE: struct<platform: string, trigger_events: list<item: string>>, POST_LURE_FINANCIAL_RISK_SIGNAL: struct<platform: string, trigger_events: list<item: string>>>>
              event_stream: list<item: struct<_event_index: string, _phase: string, event_id: string, time: timestamp[s], log_source: string, event_type: string, severity: string, entity: struct<user_id: string, device_id: string, platform: string, os_version: string, mdm_enrolled: bool, device_fingerprint_status: string>, network: struct<carrier: string, carrier_network_type: string, roaming: bool, dst_domain: string, dst_ip: string, ttl_sec: int64, domain_age_days: int64, domain_reputation: double, resolver: string, http: struct<method: string, path: string, status_code: int64>, src_ip: string>, object: struct<sender_type: string, sender_id: string, contains_url: bool, message_features: struct<urgency_language: bool, sender_verification: bool>, qtype: string, traffic_category: string, redirect_hops_total: int64, referrer_chain_domains_masked: list<item: string>, behavior: struct<signal: string, user_interaction: string, dwell_time_ms: int64, form_focus_detected: bool, tap_next: string>, user_initiated: bool, auth: struct<method: string, challenge_id: string, result: string, reason: string, burst_counter_count: int64>, risk_decision: string, result: string, fraud_score: double, flag_reason: string>, correlation: struct<incident_id: string, message_id: string, session_id: string, redirect_chain_id: string, bank_corr_id: string>, label: struct<ground_truth: string, case_type: string, attack_stage: string, mitre_mobile: string, theme: string>, original_source_hint: string, noise_metadata: struct<duplicate_of: string, reason: string>>>
              event_stream_timeline_summary: struct<total_events: int64, malicious_events: int64, benign_context_events: int64, noise_events_included: int64, phases: list<item: struct<phase: string, time_range: string, events: int64, desc: string>>, total_elapsed_minutes: double, key_time_gaps_explained: struct<Initial_Response: string, Session_Gap: string, Impact_Delay: string>>
              rule_fire_verification: struct<SMISH_REDIRECT_CHAIN: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PHISH_WARNING_BYPASS: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, SENSITIVE_FORM_SEQUENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PLATFORM_DIVERGENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, POST_LURE_FINANCIAL_RISK_SIGNAL: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>>
              scaling_template: struct<to_reach_200_events: struct<background_events_to_add: int64, background_composition: struct<messaging_sync: int64, email_sync: int64, push_notification: int64, software_update: int64, cdn_traffic: int64, utility_widgets: int64, mdm_checkin: int64, cloud_backup: int64>>>
              vs
              incident_id: string
              case_type: string
              theme: string
              difficulty_tier: string
              timeline_profile: string
              start_time: timestamp[s]
              end_time: timestamp[s]
              platform_coverage: list<item: string>
              minimum_log_sources_required: list<item: string>
              full_log_sources_used: list<item: string>
              noise_model_applied: struct<profile: string, background_event_ratio: double>
              message_samples_used: list<item: struct<target_device: string, text: string, sender_type: string, delivery_channel: string>>
              page_construction_detail_safe: struct<no_brand_copy: bool, stepper_ui_flow: list<item: string>, visual_style_notes: list<item: string>, form_risk_class: string, fields_requested: list<item: null>, detection_cues_for_human_analyst: list<item: null>>
              persona: struct<user_id: string, job_title: string, risk_profile: string, devices: list<item: struct<device_id: string, platform: string, os_version: string>>>
              behavioral_model_for_this_incident: struct<scenario: string, why_this_behavior: string, time_gap_rationale: string>
              expected_rules: struct<must_fire: list<item: null>, should_fire: list<item: null>, optional: list<item: null>, expected_fire_path: struct<>>
              event_stream: list<item: struct<_event_index: string, _phase: string, event_id: string, time: timestamp[s], log_source: string, event_type: string, severity: string, entity: struct<user_id: string, device_id: string, platform: string, os_version: string, mdm_enrolled: bool, device_fingerprint_status: string, browser: string>, network: struct<carrier: string, carrier_network_type: string, roaming: bool, dst_domain: string, dst_ip: string, ttl_sec: int64, domain_age_days: int64, domain_reputation: double, resolver: string, sni: string, tls_version: string, http: struct<method: string, path: string, status_code: int64, content_type: string, latency_ms: int64>>, object: struct<sender_type: string, sender_id: string, contains_url: bool, message_features: struct<urgency_language: bool, sender_verification: bool>, traffic_category: string, qtype: string, user_initiated: bool, posture: string, secure_dns_or_vpn: bool, os_patch_level: timestamp[s]>, correlation: struct<incident_id: string, message_id: string, session_id: string>, label: struct<ground_truth: string, case_type: string, attack_stage: null, activity_stage: string, theme: string>, original_source_hint: string, noise_metadata: struct<duplicate_of: string, reason: string>>>
              event_stream_timeline_summary: struct<total_events: int64, malicious_events: int64, benign_context_events: int64, noise_events_included: int64, phases: list<item: struct<phase: string, time_range: string, events: int64, desc: string>>, total_elapsed_minutes: double, key_time_gaps_explained: struct<sms_to_click: string, page_visit_to_end: string>>
              rule_fire_verification: struct<SMISH_REDIRECT_CHAIN: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PHISH_WARNING_BYPASS: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, SENSITIVE_FORM_SEQUENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, PLATFORM_DIVERGENCE: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>, POST_LURE_FINANCIAL_RISK_SIGNAL: struct<fires: bool, fire_path: string, trigger: string, within_window: bool>>
              scaling_template: struct<to_reach_200_events: struct<background_events_to_add: int64, background_composition: struct<messaging_sync: int64, email_sync: int64, push_notification: int64, software_update: int64, cdn_traffic: int64, utility_widgets: int64, mdm_checkin: int64, cloud_backup: int64>>>

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

license: cc-by-nc-nd-4.0 task_categories:

  • text-classification language:
  • ko tags:
  • cybersecurity
  • smishing
  • phishing
  • siem
  • mobile-security
  • synthetic-data
  • korean
  • mitre-attack
  • threat-detection
  • fraud-detection pretty_name: 2026 Korean Mobile Smishing SIEM Dataset size_categories:
  • 1K<n<10K

πŸ›‘οΈ KR-Mobile-Smishing-SIEM-Benchmark-2026

"2026λ…„ν˜• ν•œκ΅­ λͺ¨λ°”일 μŠ€λ―Έμ‹± 킬체인을 SIEM ν™˜κ²½μ—μ„œ νƒμ§€Β·ν‰κ°€ν•˜κΈ° μœ„ν•œ ν•©μ„± 벀치마크 데이터셋"

λ°°κ²½

κΈ°μ‘΄ μŠ€λ―Έμ‹± 탐지 데이터셋은 SMS ν…μŠ€νŠΈ λΆ„λ₯˜μ— ν•œμ •λ˜μ–΄, SIEM ν™˜κ²½μ—μ„œμ˜ 닀단계 곡격 탐지(λ¦¬λ””λ ‰νŠΈ 체인, 폼 νƒˆμ·¨, MFA ν”Όλ‘œ 곡격 λ“±)λ₯Ό 평가할 수 μ—†μ—ˆμŠ΅λ‹ˆλ‹€. 이 데이터셋은 SMS μˆ˜μ‹ λΆ€ν„° 금육 ν”Όν•΄ μ°¨λ‹¨κΉŒμ§€μ˜ 전체 킬체인을 SIEM 둜그 ν˜•νƒœλ‘œ μž¬ν˜„ν•©λ‹ˆλ‹€.

μ œμž‘ 방식

  • ν™”μ΄νŠΈν•΄μ»€ 페λ₯΄μ†Œλ‚˜λ₯Ό νƒ‘μž¬ν•œ AI 기반 μ‹œλ‚˜λ¦¬μ˜€ 섀계 + μ‚¬λžŒ 직접 κ²€μˆ˜ (Human-in-the-Loop)
  • MITRE ATT&CK for Mobile (v16+) ν”„λ ˆμž„μ›Œν¬ 기반 라벨링
  • 5개 상관뢄석 탐지 κ·œμΉ™μ— λŒ€ν•œ 정닡지(rule_fire_verification) 포함
  • RFC 2606 / RFC 5737 μ€€μˆ˜: λͺ¨λ“  도메인·IPΒ·μ „ν™”λ²ˆν˜Έ λΉ„ν™œμ„±ν™” 처리

πŸ“Š 데이터 νŠΉμ§• (Key Features)

Platform: iOS / Android λͺ¨λ°”일 ν™˜κ²½

Scenarios: 결제 사칭, 곡곡기관 사칭, 택배 사칭, μΈμ¦μ„œ κ°±μ‹  사칭 λ“± 10μ’…+

Case Types: malicious(곡격 성곡) / benign_lookalike(곡격 쀑단) / benign(정상) 3μ’… ν˜Όν•© β†’ True Positive Rate와 False Positive Rate λ™μ‹œ μΈ‘μ • κ°€λŠ₯

Log Sources: messaging, network, browser_security, identity, financial, noise

Labels: 이벀트 λ‹¨μœ„ ground_truth(malicious/suspicious/benign) + MITRE ATT&CK tactic/technique ID

Detection Rules: SMISH_REDIRECT_CHAIN, PHISH_WARNING_BYPASS, SENSITIVE_FORM_SEQUENCE, PLATFORM_DIVERGENCE, POST_LURE_FINANCIAL_RISK_SIGNAL β€” 각 κ·œμΉ™μ˜ λ°œλ™ μ—¬λΆ€Β·κ²½λ‘œΒ·κ·Όκ±° 포함

Noise Model: μ‹€μ œ κΈ°μ—… ν™˜κ²½μ„ λͺ¨μ‚¬ν•œ λ°°κ²½ νŠΈλž˜ν”½ 8개 μΉ΄ν…Œκ³ λ¦¬, 200 이벀트 μŠ€μΌ€μΌλ§ 지원

πŸ“Š 데이터셋 ꡬ성 (Dataset Composition)

ν•­λͺ© κ°’
Total Samples 3 (Preview) / 100+ (Full version available)
Case Types malicious, benign_lookalike, benign
Events per incident 8~15 (core) / 200 (with scaling template)
Time Range 2026
Format JSON
Language ν•œκ΅­μ–΄ (Korean) + English metadata

πŸ”’ μ•ˆμ „μ„± (Safety & Compliance)

  • λͺ¨λ“  도메인: .example TLD (RFC 2606) β€” DNS 해석 λΆˆκ°€
  • λͺ¨λ“  IP: TEST-NET λ²”μœ„ (RFC 5737) β€” λΌμš°νŒ… λΆˆκ°€
  • λͺ¨λ“  μ „ν™”λ²ˆν˜Έ: λ―Έν• λ‹Ή 번호 λŒ€μ—­
  • μ‹€μ œ λΈŒλžœλ“œλͺ…Β·λ‘œκ³ Β·UI 미포함 (no_brand_copy: true)
  • SMS 본문에 [TRAINING] μ›Œν„°λ§ˆν¬ μ‚½μž…

ν™œμš© μ˜ˆμ‹œ

  • SIEM 탐지 κ·œμΉ™μ˜ TPR / FPR 벀치마크
  • SOC 뢄석관 ν›ˆλ ¨μš© μ‹œλ‚˜λ¦¬μ˜€
  • λͺ¨λ°”일 μœ„ν˜‘ 탐지 ML λͺ¨λΈμ˜ ν•™μŠ΅ 및 평가

βœ‰οΈ Contact & License

이 λ°μ΄ν„°μ…‹μ˜ 상업적 μ΄μš©μ΄λ‚˜ 1000건 μ΄μƒμ˜ ν’€ 버전 ꡬ맀λ₯Ό μ›ν•˜μ‹œλ©΄ [senimanyc@gmail.com]둜 문의 λ°”λžλ‹ˆλ‹€.
Downloads last month
22
Total file size:
60.5 kB