Skip to main content

Module ml

Module ml 

Source
Expand description

Experimental research records for externally computed model output.

§Runtime scope

  • Build: opt-in via ml.
  • Execution: no repository runtime caller (not on any default path).
  • Default termivar scan: no.
  • Support: experimental data-model scaffold only.

The repository does not train models, cluster observations, classify anomalies, estimate exploitation success, or execute a recorded stage. Hosts may use these serializable records to carry results computed by their own reviewed systems. See docs/internals/runtime-map.md.

Structs§

AnomalyPattern
Caller-supplied anomaly-pattern record from an external model.
ClusterResult
Caller-supplied clustering record from an external model.
ExploitStage
Caller-supplied stage record containing inert host data.
ExploitationChain
Caller-supplied research-chain record. The scanner never executes its stages.
VulnerabilityPattern
Caller-supplied pattern record. No claim is inferred from this record.

Enums§

AnomalyType
Caller-assigned anomaly dimension.
MlRecordValidationError
Validation failure for externally computed research records.