cyb005-sample / README.md
pradeep-xpert's picture
Update README.md
31d9487 verified
metadata
license: cc-by-nc-4.0
task_categories:
  - tabular-classification
  - time-series-forecasting
tags:
  - cybersecurity
  - ransomware
  - threat-intelligence
  - apt
  - synthetic-data
  - double-extortion
  - backup-recovery
  - mitre-attack
  - incident-response
  - raas
pretty_name: CYB005  Synthetic Ransomware Attack Simulation (Sample)
size_categories:
  - 10K<n<100K

CYB005 — Synthetic Ransomware Attack Simulation Dataset (Sample)

XpertSystems.ai Synthetic Data Platform · SKU: CYB005-SAMPLE · Version 1.0.0

This is a free preview of the full CYB005 — Synthetic Ransomware Attack Simulation Dataset product. It contains roughly ~10% of the full dataset at identical schema, actor-tier distribution, and statistical fingerprint, so you can evaluate fit before licensing the full product.

🤖 Trained baseline available: xpertsystems/cyb005-baseline-classifier — XGBoost + PyTorch MLP for 4-tier threat-actor attribution (the README's stated headline use case), group-aware split by campaign, multi-seed evaluation (ROC-AUC 0.853 ± 0.031), honest leakage audit of every per-timestep feature.

Note: This sample is intentionally larger than the other CYB SKU samples. CYB005 benchmarks are conditional on small actor-tier subsets (e.g. nation_state campaigns are ~10% of the fleet), so a larger sample is needed to demonstrate the full product's benchmark calibration reliably.

File Rows (sample) Rows (full) Description
victim_topology.csv ~300 ~3,200 Network segment registry
campaign_summary.csv ~500 ~5,500 Per-campaign outcome aggregates
campaign_events.csv ~190,137 ~60,000 Discrete campaign event log
attack_timelines.csv ~37,489 ~290,000 Per-timestep campaign trajectory data

Dataset Summary

CYB005 simulates end-to-end ransomware campaign lifecycles as a 7-phase state machine across enterprise, cloud, and OT/ICS environments, with:

  • 4 actor capability tiers: lone_actor, organised_syndicate, raas_affiliate, nation_state_nexus — with per-tier encryption speed, ransom demand distributions, wiper component probabilities, and lateral movement aggression
  • 6 victim backup maturity tiers: no_backup, local_only, network_attached, cloud_replicated, immutable_object_lock, air_gapped_gold_standard — with empirically-calibrated recovery probabilities
  • 8 segment types: corporate_lan, dmz, cloud_workload, ot_ics_control, endpoint_subnet, soc_management, zero_trust_zone, backup_repository
  • 7 attack phases: initial_access, persistence, privilege_escalation, lateral_movement, data_exfiltration, encryption_deployment, ransom_demand
  • Double extortion modeling (data exfiltration + encryption)
  • VSS (Volume Shadow Copy) deletion, wiper components, and worm spread
  • Living-off-the-Land (LotL) abuse and EDR signature lag modeling
  • Financial impact scoring with ransom demand × payment probability

Trained Baseline Available

A working baseline classifier trained on this sample is published at xpertsystems/cyb005-baseline-classifier.

Component Detail
Task 4-class threat-actor capability-tier attribution (the README's headline use case)
Models XGBoost (model_xgb.json) + PyTorch MLP (model_mlp.safetensors)
Features 63 (after one-hot encoding); pipeline included as feature_engineering.py
Split Group-aware by campaign_id — train/val/test campaigns disjoint
Validation Single seed + multi-seed aggregate across 10 seeds
Demo inference_example.ipynb — end-to-end copy-paste
Headline metrics XGBoost: accuracy 0.603 ± 0.040, macro ROC-AUC 0.853 ± 0.031 (multi-seed)

This is the first XpertSystems baseline to ship the dataset's stated headline use case (rather than pivoting to a phase-prediction subtask as the smaller CYB002 / CYB003 / CYB004 samples required). CYB005's 500-campaign sample is large enough that tier attribution learns honestly under group-aware splitting.

Calibrated Benchmark Targets

The full product is calibrated to 12 benchmark metrics drawn from authoritative ransomware threat intelligence sources (Mandiant M-Trends, CrowdStrike GTR, Coveware Quarterly Ransomware Report, Sophos State of Ransomware, IBM CODB, Verizon DBIR, CISA #StopRansomware, Chainalysis). The sample preserves the same calibration:

Test Target Observed Verdict
01_blast_radius_pct_organised_syndicate_low_seg 0.3700 0.3302 ✓ PASS
02_dwell_time_pre_detonation_hrs_median 204.0000 226.1000 ✓ PASS
03_ransom_paid_rate_all_tiers 0.2900 0.2941 ✓ PASS
04_recovery_without_payment_rate_immutable 0.7200 0.7292 ✓ PASS
05_double_extortion_rate_raas_syndicate 0.7700 0.7400 ✓ PASS
06_mttd_hrs_global_median 192.0000 203.5600 ✓ PASS
07_ransom_demand_usd_median_raas 650,000 633,445 ✓ PASS
08_vss_deletion_success_rate 0.6800 0.6529 ✓ PASS
09_edr_alert_rate_per_lateral_move 0.5400 0.5123 ✓ PASS
10_wiper_component_rate_nation_state 0.2200 0.2933 ~ MARGINAL
11_backup_destruction_rate_weak_tiers 0.4200 0.4126 ✓ PASS
12_financial_impact_score_syndicate 0.6100 0.5810 ✓ PASS

Note: some benchmarks (e.g. wiper component rate, blast radius) require larger sample sizes to converge tightly because they're conditional on small-population subsets (e.g. nation-state campaigns are ~10% of fleet). The full product passes all 12 benchmarks at Grade A+ or better.

Schema Highlights

attack_timelines.csv (primary file, per-timestep)

Column Type Description
campaign_id string Unique campaign identifier
actor_id string Threat actor ID
timestep int Step in 7-phase lifecycle (0–74)
campaign_phase string 1 of 7 phases
actor_capability_tier string lone_actor / organised_syndicate / raas_affiliate / nation_state_nexus
segment_id string FK to victim_topology.csv
backup_maturity_tier string 6 tiers from no_backup to air_gapped
endpoints_compromised int Cumulative endpoints affected
blast_radius_pct float Fleet-wide compromise percentage
lateral_pivots int Lateral movement count
edr_alerted int Boolean — EDR alert raised
siem_correlated int Boolean — SIEM correlation event
lotl_technique_used string LotL binary if any
vss_deletion_attempted int Boolean — Volume Shadow Copy deletion
wiper_component_deployed int Boolean — destructive wiper present
data_exfiltrated_gb float Cumulative exfiltrated data
dwell_hours float Cumulative attacker dwell time
c2_beacon_active int C2 channel beaconing flag

campaign_summary.csv (per-campaign outcome)

Column Type Description
campaign_id, actor_id string Identifiers
actor_capability_tier string Tier classification target
backup_maturity_tier string Victim backup posture
campaign_outcome string success / partial / detected / aborted
ransom_demand_usd float Ransom amount demanded
ransom_paid_flag int Boolean — ransom paid
recovery_without_payment_flag int Boolean — restored from backup
double_extortion_flag int Boolean — data leak threat
wiper_component_flag int Boolean — wiper deployed
dwell_time_pre_detonation_hrs float Hours from access to encryption
mean_time_to_detect_hrs float Hours from access to first detection
financial_impact_score float Composite impact score (0–1)
blast_radius_pct float Fleet compromise percentage

See campaign_events.csv and victim_topology.csv for the discrete event log and segment registry schemas respectively.

Suggested Use Cases

  • Training ransomware classifier models — worked example available
  • Backup posture risk modeling — predict recovery likelihood from 6-tier backup maturity
  • Dwell time forecasting under varying actor capability and defender maturity
  • Double extortion prediction (data theft + encryption modeling)
  • Wiper component detection — distinguishing destructive vs financial ransomware
  • VSS deletion / shadow copy abuse detection
  • Financial impact estimation — ransom demand + payment probability
  • EDR alert correlation — SIEM signal-to-noise modeling
  • Incident response simulation — purple-team exercises with calibrated attacker behavior

Loading the Data

import pandas as pd

timelines  = pd.read_csv("attack_timelines.csv")
summaries  = pd.read_csv("campaign_summary.csv")
events     = pd.read_csv("campaign_events.csv")
topology   = pd.read_csv("victim_topology.csv")

# Join per-timestep data with campaign-level labels and topology
enriched = timelines.merge(summaries, on=["campaign_id", "actor_id"], how="left",
                           suffixes=("", "_summary"))
enriched = enriched.merge(topology, on="segment_id", how="left")

# Actor-tier classification target
y_tier = summaries["actor_capability_tier"]

# Binary outcomes
y_paid   = summaries["ransom_paid_flag"]
y_recovered = summaries["recovery_without_payment_flag"]
y_wiper  = summaries["wiper_component_flag"]

For a worked end-to-end example with actor-tier classification, group-aware splitting, and feature engineering, see the inference notebook in the baseline classifier repo.

License

This sample is released under CC-BY-NC-4.0 (free for non-commercial research and evaluation). The full production dataset is licensed commercially — contact XpertSystems.ai for licensing terms.

Full Product

The full CYB005 dataset includes ~358,000 rows across all four files, with calibrated benchmark validation against 12 metrics drawn from authoritative ransomware threat intelligence sources.

📧 pradeep@xpertsystems.ai 🌐 https://xpertsystems.ai

Citation

@dataset{xpertsystems_cyb005_sample_2026,
  title  = {CYB005: Synthetic Ransomware Attack Simulation Dataset (Sample)},
  author = {XpertSystems.ai},
  year   = {2026},
  url    = {https://huggingface.co/datasets/xpertsystems/cyb005-sample}
}

Generation Details

  • Generator version : 1.0.0
  • Random seed : 42
  • Generated : 2026-05-16 14:03:22 UTC
  • Campaign model : 7-phase ransomware kill-chain state machine
  • Overall benchmark : 97.7 / 100 (grade A+)