Watchlists, Feeds, Reports, and IOCs

Watchlists are a powerful feature of Carbon Black Cloud Enterprise EDR. They allow an organization to set-and-forget searches on their endpoints’ incoming events data, providing the administrator the opportunity to sift through high volumes of activity and focus attention on those that matter.

Note: Use of these APIs requires that the organization be enabled for Enterprise EDR. Verify this by logging into the Carbon Black Cloud Console, opening the menu in the upper right corner, and checking for an ENABLED flag against the “Enterprise EDR” entry.

All examples here assume that a Carbon Black Cloud SDK connection has been set up, such as with the following code:

>>> from cbc_sdk import CBCloudAPI
>>> api = CBCloudAPI(profile='sample')

Setting up a connection is documented here: Getting Started with the Carbon Black Cloud Python SDK - “Hello CBC”

About the Objects

An indicator of compromise (IOC) is a query, list of strings, or list of regular expressions which constitutes actionable threat intelligence that the Carbon Black Cloud is set up to watch for. Any activity that matches one of these may indicate a compromise of an endpoint.

A report groups one or more IOCs together, which may reflect a number of possible conditions to look for, or a number of conditions related to a particular target program or type of malware. Reports can be used to organize IOCs.

A watchlist contains reports (either directly or through a feed) that the Carbon Black Cloud is matching against events coming from the endpoints. A positive match will trigger a “hit,” which may be logged or result in an alert.

A feed contains reports which have been gathered by a single source. They resemble “potential watchlists.” A watchlist may be easily subscribed to a feed, so that any reports in the feed act as if they were in the watchlist itself, triggering logs or alerts as appropriate.

Setting Up a Basic Custom Watchlist

Creating a custom watchlist that can watch incoming events and/or generate alerts requires three steps:

  1. Create a report including one or more Indicators of Compromise (IOCs).

  2. Add that report to a watchlist.

  3. Enable alerting on the watchlist.

Creating a Report

In this example, a report is created, adding one or more IOCs to it:

>>> from cbc_sdk.enterprise_edr import Report, IOC_V2
>>> builder = Report.create(api, "Unsigned Browsers", "Unsigned processes impersonating browsers", 5)
>>> builder.add_tag("compliance").add_tag("unsigned_browsers")
>>> builder.add_ioc(IOC_V2.create_query(api, "unsigned-chrome",
...                 "process_name:chrome.exe NOT process_publisher_state:FILE_SIGNATURE_STATE_SIGNED"))
>>> report =
>>> report.save_watchlist()

Reports should always be given a title that’s sufficiently unique within your organization, so as to minimize the chances of confusing two or more Reports with each other. Carbon Black Cloud will generate unique id values for each report, but does not enforce any uniqueness constraint on the title of reports.

Alternatively, you can update an existing report, adding more IOCs and/or replacing existing ones. To find an existing report associated with a watchlist, you must look in the watchlist’s reports collection:

>>> from cbc_sdk.enterprise_edr import Watchlist, Report, IOC_V2
>>> watchlist =, 'R4cMgFIhRaakgk749MRr6Q')
>>> report_list = [report for report in watchlist.reports if == '47474d40-1f94-4995-b6d9-1d1eea3528b3']
>>> report = report_list[0]
>>> report.append_iocs([IOC_V2.create_query(api, 'evil-connect', 'netconn_ipv4:')])
>>> report.update()

Adding the Report to a Watchlist

Now, add the new Report to a new Watchlist:

>>> from cbc_sdk.enterprise_edr import Watchlist
>>> builder = Watchlist.create(api, "Suspicious Applications")
>>> builder.set_description("Any signs of suspicious applications running on endpoints").add_reports([report])
>>> watchlist =

If you already have an existing Watchlist you wish to enhance, you can add Reports to the existing Watchlist:

>>> # "report" contains the Report that was created in the previous example
>>> from cbc_sdk.enterprise_edr import Watchlist
>>> watchlist ='Watchlist', 'R4cMgFIhRaakgk749MRr6Q')
>>> watchlist.add_reports([report])

Enabling Alerting on a Watchlist

When either the alerts_enabled or tags_enabled attributes of a watchlist are True, that Watchlist will create data you can act on - either alerts or hits, respectively; if both are False, the Watchlist is effectively disabled.

Once you have the Watchlist configured with the IOCs that are generating the kinds of hits (results) you are after, you can enable Alerting for the Watchlist, which will allow matches against the reports in the watchlist to generate alerts. If a watchlist identifies suspicious behavior and known threats in your environment, you will want to enable alerts to advise you of situations where you may need to take action or modify policies.

>>> watchlist.enable_alerts()

A Closer Look at IOCs

In this document, only the “v2” IOCs are covered; the “v1” IOCs are only provided for backwards compatibility reasons. They are officially deprecated, and are converted, internally, to this type.

IOCs can be classified into two general types, depending on their match_type value:

Query IOCs are those with a match_type of query; their values_list contains a single string that specifies a query compatible with process searches. For example, the following IOC looks for the process git.exe that does not connect to one of a specified list of IP addresses:

    "id": "example_1",
    "match_type": "query",
    "values": ["process_name:git.exe NOT (netconn_ipv4: OR netconn_ipv4:
                OR netconn_ipv4: OR netconn_ipv4:"]

Query IOCs must always use field-prefixed queries (key-value pairs); they do not support just searching for a value without a field specified. Values in query clauses that do not specify fields will be ignored.


process_name:chrome.exe AND


process_name:chrome.exe AND netconn_ipv4:

Query IOCs may search on CIDR address ranges, e.g.: netconn_ipv4:

Query IOCs are searched every 5 minutes by the Carbon Black Cloud, and are tested against a rolling window of the last hour’s worth of data for the organization. (They will not generate hits or alerts for process attributes that were reported more than an hour in the past.) They may employ any searchable field as documented here, and may employ complex query logic.

Ingress IOCs are those with a match_type of equality or regex; they use the field element to specify the name of a field to examine the value of, and the values_list element to specify a list of values to match against (in the case of match_type being equality) or regular expressions to match against (in the case of match_type being regex). For example, this IOC will match any process that initiates a connection to one of two listed IP addresses:

    "id": "example_2",
    "match_type": "equality",
    "field": "netconn_ipv4",
    "values": ["", ""]

This IOC will match any process running with an executable name beginning with “quake”:

    "id": "example_3",
    "match_type": "regex",
    "field": "process_name",
    "values": ["quake.*\\.exe"]

(Note the use of the backslash to escape the ‘.’ that separates the file extension from the name. It must be doubled to escape it in Python itself.)

Ingress IOCs are searched as soon as the data is received from any endpoint, and may use any process field (as documented here; the fields that may be used in this context are tagged with PROCESS) in their field element, whether searchable or not. For the searches they are capable of, they are more efficient than query IOCs, and also easier to add additional search target values to. They can, however, only search on a single field at a time.

Note: Ingress IOCs cannot be edited in the Carbon Black Cloud console UI at this time, due to a UI limitation on editing two properties of an IOC at the same time.

You can include more than one entry (query or match element) in an individual IOC, but in order to ignore or disable one of those entries, you would either have to edit the IOC or disable it entirely (thus disabling all entries in that IOC). It is recommended to use only one entry per IOC, for ease of management, unless you have already vetted the entries and don’t expect to have to disable them individually.

Both IOCs and reports may include a link property, which is used by the Carbon Black Cloud console UI as a hint to indicate that this IOC or report is being managed outside of the console. If this property is not None, the console UI will disable the ability to edit the IOC or report, but they can still be edited via the API.

Creating an IOC

You can create an IOC via the IOC_V2 class, there are 3 avaliable methods that you can use to initiate your IOC: IOC_V2.create_query, IOC_V2.create_equality, IOC_V2.create_regex.

Creating an equality IOC

>> from cbc_sdk import CBCloudAPI
>> from cbc_sdk.enterprise_edr import IOC_V2
>> cbcsdk = CBCloudAPI(profile="default")
>> IOC_V2.create_equality(cbcsdk, None, "netconn_domain", ["localhost.local"])
<cbc_sdk.enterprise_edr.threat_intelligence.IOC_V2: id ad361179-d586-4c99-af3e-821224cc0fd9> @ https://<CBCInstanceURL>

Creating a query IOC

>> IOC_V2.create_query(cbcsdk, None, "{process_hash:098f6bcd4621d373cade4e832627b4f6}")
<cbc_sdk.enterprise_edr.threat_intelligence.IOC_V2: id 36d68cab-4739-4aa6-afcc-2921d2e5573e> @ https://<CBCInstanceURL>

Creating a regex IOC

>> IOC_V2.create_regex(cbcsdk, None, "process_name", r"(^/usr/.*$)|(^/bin/.*$)")
<cbc_sdk.enterprise_edr.threat_intelligence.IOC_V2: id 5170a04c-bbfc-4449-b939-d5fc9f55d555> @ https://<CBCInstanceURL>

Removing and adding an IOC from a Report

If you want to remove an IOC from a report, you will need the IOC id and the report id.

>> from cbc_sdk.enterprise_edr import Report
>> ioc_id = "<ioc_id>"
>> report ="<report_id>", feed_id="<feed_id>")[0]
<cbc_sdk.enterprise_edr.threat_intelligence.Report: id 1e69c54e-7cc9-41b8-9d1d-3fd59a003d8a> @ https://<CBCInstanceURL>
>> report.remove_iocs_by_id([ioc_id])
>> report.update()
<cbc_sdk.enterprise_edr.threat_intelligence.Report: id 1e69c54e-7cc9-41b8-9d1d-3fd59a003d8b> @ https://<CBCInstanceURL> (*)

Adding the IOC into the report works the same way:

>> from cbc_sdk.enterprise_edr import Report, IOC_V2
>> ioc_id = "<ioc_id>"
>> report ="<report_id>", feed_id="<feed_id>")[0]
<cbc_sdk.enterprise_edr.threat_intelligence.Report: id 1e69c54e-7cc9-41b8-9d1d-3fd59a003d8a> @ https://<CBCInstanceURL>
>> ioc = IOC_V2.create_regex(cbcsdk, None, "process_name", r"(^/usr/.*$)|(^/bin/.*$)")
>> report.append_iocs([ioc])
>> report.update()
<cbc_sdk.enterprise_edr.threat_intelligence.Report: id 1e69c54e-7cc9-41b8-9d1d-3fd59a003d8b> @ https://<CBCInstanceURL> (*)


Calling the method after the insertion or removal of IOC does not update the report and it’s likely to result in a bad call to the API.

If the report is in a watchlist instead of a feed then you have to get the appropriate watchlist and iterate over the reports.

>> from cbc_sdk.enterprise_edr import Watchlist, Report, IOC_V2
>> ioc_id = "<ioc_id>"
>> report_id = "<report_id>"
>> watchlist =, "<watchlist_id>")
<cbc_sdk.enterprise_edr.threat_intelligence.Watchlist: id <watchlist_id>> @ https://<CBCInstanceURL>
>> ioc = IOC_V2.create_regex(cbcsdk, None, "process_name", r"(^/usr/.*$)|(^/bin/.*$)")
>> reports = watchlist.reports
>> report = [report_ for report_ in reports if == report_id][0]
>> report.append_iocs([ioc])
>> report.update()
<cbc_sdk.enterprise_edr.threat_intelligence.Report: id 1e69c54e-7cc9-41b8-9d1d-3fd59a003d8b> @ https://<CBCInstanceURL> (*)

Tips for Using IOCs

  • You can safely ignore certain fields in an IOC. For example, fields like alert_id and process_guid will always uniquely identify just a single record in your organization’s data, whereas a field like org_id will be a constant across all your organization’s data.

  • Timestamp fields such as backend_timestamp are useful in ad-hoc queries, to look for data occurring before or after a certain date, but are of limited usefulness over the span of time a watchlist may be running.

  • A list of hashes (such as with process_sha256) can be of limited value. They are inconvenient to keep current, especially as software (whether legitimate or malicious) gets updated over time, but are definitely easier to manage with equality IOCs.

  • Counter fields (such as netconn_count) can be useful with range queries to locate processes that are using a large number of resources. For example, the query netconn_count:[500 TO *] will match only processes that make a large number of network connections.

  • When using ingress IOCs, be careful of errant characters in the values list, such as leading or trailing whitespace or embedded newline characters. These errant characters may cause the IOCs to fail to match, leading to false negative results.

  • equality IOCs for IPv4 fields (e.g. netconn_remote_ipv4) cannot support CIDR notation; full IP addresses must be used.

  • equality IOCs for IPv6 fields (e.g. netconn_remote_ipv6) do not support standard or CIDR notation at this time. All IPv6 addresses must omit colon characters, spell out all zeroes in the address, and represent all alphabetic characters in uppercase. For example, “ff02::fb” becomes “FF0200000000000000000000000000FB”.


Another way of managing reports is to attach them to a feed. Feeds can contain multiple reports, and a feed can be attached to a watchlist, effectively making the contents of the watchlist equivalent to the contents of the feed.

Feeds are in effect “potentially-subscribable Watchlists”. A Feed has no effect on your organization until it is subscribed to, by creating a Watchlist containing that feed. Once subscribed (and until it’s disabled or unsubscribed), a watchlist will generate hits (and alerts if you have enabled them) for any matches against any of the IOCs in any of that feed’s enabled reports.

Note: The feeds that are created by these examples are private feeds, meaning they are only visible within an organization and can be created by anyone with sufficient privileges in the organization. There are additional types of feeds; reserved feeds can only be created by MSSPs, and public feeds can only be created or edited by VMware Carbon Black.

A new feed may be created as follows (assuming the new report for that feed is stored in the report variable):

>>> from cbc_sdk.enterprise_edr import Feed
>>> builder = Feed.create(api, 'Suspicious Applications', '',
...                       'Any signs of suspicious applications running on our endpoints', 'external_threat_intel')
>>> builder.set_source_label('Where the info is coming from')
>>> builder.add_reports([report])
>>> feed =

If you have an existing feed, a new report may be added to it as follows (assuming the new report is stored in the report variable):

>>> from cbc_sdk.enterprise_edr import Feed
>>> feed.append_reports([report])

To update or delete an existing report in a feed, look for it in the feed’s reports collection, then call the update() method on the report to replace its contents, or the delete() method on the report to delete it entirely. The replace_reports() method on the Feed object may also be used, but caution must be taken, as that method will replace all of the reports in a feed at once.

To subscribe to a feed, a new watchlist must be created around it:

>>> watchlist = Watchlist.create_from_feed(feed, "Subscribed feed", "Subscription to the new feed")

Limitations of Reports and Watchlists

Individual reports may contain no more than 10,000 IOCs. Reports containing more than 1,000 IOCs will not be editable via the Carbon Black Cloud console UI, but may still be managed using APIs.

Individual watchlists may contain no more than 10,000 reports. Any more than that may lead to timeouts when managing the watchlist through the Carbon Black Cloud console UI, and possibly when managing it through APIs as well.