For the complete documentation index, see llms.txt. This page is also available as Markdown.
Cortex XSIAM

Disable indicator extraction for scripts or integrations

Prevent selected scripts or integrations from automatically extracting indicators.

By default, system-wide indicator extraction and enrichment is disabled.

If you have the TIM add-on and you have enabled system-wide indicator extraction and enrichment, the procedure below enables you to disable indicator extraction for a specific script or integration.

Note

The TIM add-on is included in the Cortex XSIAM Premium license.

  • To disable indicator extraction for a script, add the IgnoreAutoExtract entry with the value of true, when returning an entry.

    For example:

    entry = {
        'Type': entryTypes['note'],
        'Contents': {
        'Echo' : demisto.args()['echo']
            },
        'ContentsFormat': formats['json'],
        'ReadableContentsFormat': formats['markdown'],
        'HumanReadable': hr,
        'IgnoreAutoExtract' : True
       }
  • To disable indicator extraction for an integration, add the 'IgnoreAutoExtract' entry with the value of true, when returning an entry.

    For example in the ServiceNow integration:

    entry = {
            'Type': entryTypes['note'],
            'Contents': result,
            'ContentsFormat': formats['json'],
            'ReadableContentsFormat': formats['markdown'],
            'HumanReadable': tableToMarkdown('ServiceNow ticket', hr, headers=headers, removeNull=True),
            'EntryContext': {
                'Ticket(val.ID===obj.ID)': context,
                'ServiceNow.Ticket(val.ID===obj.ID)': context
            },
            'IgnoreAutoExtract': True
        }
        entries.append(entry)
        return entries

For more information about command results in Python, see Python code conventions for CommandResults.

Last updated

Was this helpful?