Locy Cyber Exposure Twin Dataset (Flagship #3)

This dataset is a deterministic, documentation-friendly snapshot for the Cyber Exposure-to-Remediation Decision Twin notebook.

What it contains

• assets.csv: OT/IT asset inventory with business criticality and internet exposure.
• vulnerabilities.csv: CVE metadata (severity, family, attack surface).
• kev_snapshot.csv: curated KEV flags.
• epss_snapshot.csv: curated EPSS probabilities.
• vuln_findings.csv: asset-to-CVE findings and SLA metadata.
• asset_dependencies.csv: directed dependency graph for blast propagation.
• remediation_actions.csv: candidate remediation actions by CVE.
• knowledge_docs.csv: remediation/advisory corpus with text + small demo embeddings.
• notebook_cases.csv: deterministic focus assets for notebook narrative.
• manifest.json: generation metadata and source references.

Provenance

The snapshot is inspired by public cyber sources and normalized for reproducible docs/CI execution:

• CISA KEV Catalog
• FIRST EPSS
• NVD CVE records
• MITRE ATT&CK references

See manifest.json for URLs and snapshot metadata.

Regeneration

From repository root:

uv run python website/scripts/prepare_cyber_exposure_twin_data.py

This rewrites all CSVs and manifest.json deterministically.