PostExploitBench: Multi-Host Post-Exploitation Cyber Range Benchmark

AgentCyberRange created PostExploitBench, a benchmark for multi-host post-exploitation tasks. The complete dataset contains 8 self-contained cyber ranges modeling enterprise-like environments with 156 target hosts. The dataset was last updated on June 16, 2026.

Use Cases

• Evaluating autonomous agent performance on multi-stage compromise tasks based on the described cyber range structure.
• Benchmarking lateral movement and privilege escalation algorithms based on the presence of pivot hosts and vulnerable targets.
• Testing intrusion detection systems against realistic attack scenarios based on the inclusion of decoys and supporting services.

Strengths

• Contains 8 self-contained cyber ranges, providing multiple distinct test environments.
• Includes 156 target hosts, offering a substantial scale for evaluation.
• Models enterprise-like environments with components such as entry services, internal networks, and decoys.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

AgentCyberRange

Collection Method

Likely constructed as simulated cyber ranges for benchmarking.

Freshness

Last updated 2026-06-16 12:12:15; freshness should be verified.