TypeSafe Jev
This Integration is part of the TypeSafe Jev Pack.#
Supported versions
Available on Cortex XSOAR (versions 6.10.0 and later) and Cortex XSIAM.
Use TypeSafe Jev typed judgments and calibrated probabilities in Cortex XSIAM and XSOAR playbooks.
Configure TypeSafe Jev in Cortex#
| Parameter | Description | Required |
|---|---|---|
| Server URL | True | |
| API Key | The TypeSafe API key used to authenticate Jev requests. Generate or manage keys in the TypeSafe console. | True |
| Model | The default Jev model. Use a pinned version when playbook thresholds have been validated against that version. | False |
| Model (optional free text override) | The optional model name that overrides the selected model. | False |
| Request timeout | The HTTP request timeout in seconds. | True |
| Trust any certificate (not secure) | False | |
| Use system proxy settings | False |
Commands#
You can execute these commands from the CLI, as part of an automation, or in a playbook. After you successfully execute a command, a DBot message appears in the War Room with the command details.
jev-evaluate#
Evaluate a state against one or more typed TypeSafe questions in a single request.
Base Command#
jev-evaluate
Input#
| Argument Name | Description | Required |
|---|---|---|
| state | The text or JSON state to evaluate. Valid JSON is sent as structured state; other values are sent as text. | Required |
| questions | The JSON object whose keys are question IDs and whose values are TypeSafe Noul, Choice, or Score question definitions. | Required |
| model | The optional model override for this request. | Optional |
Context Output#
| Path | Type | Description |
|---|---|---|
| TypeSafeJev.Evaluation.model | String | The versioned model that handled the request. |
| TypeSafeJev.Evaluation.answers | Unknown | The typed answers keyed by the submitted question IDs. |
| TypeSafeJev.Evaluation.usage.input_tokens | Number | The number of input tokens used by the request. |
| TypeSafeJev.Evaluation.usage.output_tokens | Number | The number of output tokens reported for the request. |
jev-noul#
Evaluate one yes/no judgment and return the probability that the answer is yes.
Base Command#
jev-noul
Input#
| Argument Name | Description | Required |
|---|---|---|
| state | The text or JSON state to evaluate. | Required |
| instructions | The narrow yes/no judgment to make about the state. | Required |
| true_criteria | The optional description of what a yes result means. | Optional |
| false_criteria | The optional description of what a no result means. | Optional |
| model | The optional model override for this request. | Optional |
Context Output#
| Path | Type | Description |
|---|---|---|
| TypeSafeJev.Noul.model | String | The versioned model that handled the request. |
| TypeSafeJev.Noul.probability | Number | The probability of yes, from 0 to 1. |
| TypeSafeJev.Noul.usage | Unknown | The token usage for the request. |
jev-choice#
Select one option from a defined set and return the complete probability distribution.
Base Command#
jev-choice
Input#
| Argument Name | Description | Required |
|---|---|---|
| state | The text or JSON state to evaluate. | Required |
| instructions | The decision Jev should make about the state. | Required |
| criteria | The JSON object mapping each option name to its description or null. Include a no-match option when appropriate. | Required |
| model | The optional model override for this request. | Optional |
Context Output#
| Path | Type | Description |
|---|---|---|
| TypeSafeJev.Choice.model | String | The versioned model that handled the request. |
| TypeSafeJev.Choice.choice | String | The selected option. |
| TypeSafeJev.Choice.probabilities | Unknown | The probability distribution over all supplied options. |
| TypeSafeJev.Choice.confidence | Number | The confidence derived from the probability distribution. |
| TypeSafeJev.Choice.usage | Unknown | The token usage for the request. |
jev-score#
Score state against ordered levels and return the weighted score and probability distribution.
Base Command#
jev-score
Input#
| Argument Name | Description | Required |
|---|---|---|
| state | The text or JSON state to evaluate. | Required |
| instructions | The dimension Jev should score about the state. | Required |
| criteria | The JSON array of two to ten ordered, standalone level descriptions. | Required |
| model | The optional model override for this request. | Optional |
Context Output#
| Path | Type | Description |
|---|---|---|
| TypeSafeJev.Score.model | String | The versioned model that handled the request. |
| TypeSafeJev.Score.score | Number | The probability-weighted score across the ordered levels. |
| TypeSafeJev.Score.legend | Unknown | The mapping from numeric level to its description. |
| TypeSafeJev.Score.probabilities | Unknown | The probability distribution over the levels. |
| TypeSafeJev.Score.confidence | Number | The confidence derived from the probability distribution. |
| TypeSafeJev.Score.usage | Unknown | The token usage for the request. |
jev-list-models#
List model aliases available to the configured TypeSafe account.
Base Command#
jev-list-models
Input#
| Argument Name | Description | Required |
|---|
Context Output#
| Path | Type | Description |
|---|---|---|
| TypeSafeJev.Model.name | String | The model name or alias. |
| TypeSafeJev.Model.description | String | The model description. |
| TypeSafeJev.Model.release_date | Date | The model release date. |