Skip to content

CRML CLI Commands

The crml command-line interface provides a thin wrapper around the Python runtime.

crml validate

Validate a CRML file against the JSON schema.

crml validate model.yaml
  • Exits with code 0 on success
  • Raises a JSON schema validation error on failure

crml simulate

Run a Monte Carlo simulation based on the CRML model.

crml simulate model.yaml -n 30000

Options:

  • -n, --runs (int): number of Monte Carlo runs (default: 10000)
  • -s, --seed (int): random seed for reproducibility
  • -f, --format (text|json): output format (default: text)
  • --fx-config (path): path to FX configuration file for currency conversion

Currency Handling

Models should specify currency on monetary values:

severity:
  model: lognormal
  parameters:
    median: "100 000"
    currency: EUR
    sigma: 1.2

To convert currencies and control output display, use an FX config file:

crml simulate model.yaml --fx-config fx-config.yaml

Example fx-config.yaml:

base_currency: USD
output_currency: EUR  # Display results in Euros
as_of: "2025-01-15"
rates:
  EUR: 1.08
  GBP: 1.26
  JPY: 0.0066

Output example:

==================================================
CRML Simulation Results
==================================================
Model: data-breach-simple
Runs: 10,000
Runtime: 45.23 ms
Currency: EUR (€)

==================================================
Risk Metrics
==================================================
EAL (Expected Annual Loss):  €27,100,000.00
VaR 95%:                      €39,700,000.00
VaR 99%:                      €44,300,000.00
VaR 99.9%:                    €68,400,000.00
==================================================

crml run (legacy)

Alias for crml simulate with text output.

crml run model.yaml --runs 30000

crml explain

Print a short human-readable summary of the model.

crml explain model.yaml

Example:

CRML Model: financial-services-risk-model
Description: Unified enterprise cyber risk model.
Assets: 18000
Frequency model: gamma_poisson
Severity model: mixture

Error handling

  • Invalid YAML/JSON → loader error
  • Schema mismatch → validation error
  • Unknown model names → runtime NotImplementedError