python-act is a library used to connect to the ACT platform.
The source code for this API is availble on github and on PyPi.
Install from PyPi:
$ pip3 install act-api
The platform has a REST api, and the goal of this library is to expose all functionality in the API.
The act platform is built on two basic types, the object and fact.
Objects are universal elements that can be referenced uniquely by its value. An example of an object can be an IP address.
Facts are assertions or obsersvations that ties objects together. A fact may or may not have a value desribing further the fact.
Facts can be linked on or more objects. Below, the seenIn fact is linked to both an ipv4 object and report object, but the hasTitle fact is only linked to a report.
| Object type | Object value | Fact type | Fact value | Object type | Object value |
|---|---|---|---|---|---|
| ipv4 | 127.0.0.1 | seenIn | report | report | cbc80bb5c0c0f8944bf73(...) |
| report | cbc80bb5c0c0f8944bf73(...) | hasTitle | Threat Intel Summary | n/a | n/a |
- Most functions returns an object that can be chained
- Attributes can be accessed using dot notation (e.g fact.name and fact.type.name)
Connct to the API using an URL where the API is exposed and a user ID:
>>> import act
>>> c = act.api.Act("https://act-eu1.mnemonic.no", user_id = 1, log_level = "warning")
The returned object exposes most of the API in the ACT platform:
- fact - Manage facts
- fact_search - Search facts
- fact_type - Instantiate a fact type
- get_fact_types - Get fact types
- object - Manage objects
- object_search - Searh objects
- origin - Manage origins
- get_object_types - Get object types
Additional arguments to act.api.Act can be passed on to requests using the requests_common_kwargs, which mans you can add for instance auth if the instance is behind a reverse proxy with HTTP authentication:
>>> c = act.api.Act("https://act-eu1.mnemonic.no", user_id = 1, log_level = "warning", requests_common_kwargs = {"auth": ("act", "<PASSWORD>")})
Create a fact by calling fact(). The result can be chained using one or more source(), destination() or bidirectionial() to add linked objects.
>>> f = c.fact("seenIn", "report").source("ipv4", "127.0.0.1").destination("report", "87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7")
>>> f
Fact(type='seenIn', value='report', source_object=Object(type='ipv4', value='127.0.0.1'), destination_object=Object(type='report', value='87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7'))
The fact is not yet added to the platform. User serialize() or json() to see the parameters that will be sent to the platform when the fact is added.
>>> f.serialize()
{'type': 'seenIn', 'destinationObject': {'type': 'report', 'value': '87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7'}, 'accessMode': 'Public', 'value': 'report', 'sourceObject': {'type': 'ipv4', 'value': '127.0.0.1'}, 'bidirectionalBinding': False}
>>> f.json()
'{"type": "seenIn", "destinationObject": {"type": "report", "value": "87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7"}, "accessMode": "Public", "value": "report", "sourceObject": {"type": "ipv4", "value": "127.0.0.1"}, "bidirectionalBinding": false}'<Paste>
Since the fact is not yet added it does not have an id.
>>> print(f.id)
None
Use add() to add the fact to the platform.
>>> f.add()
Fact(type='seenIn', value='report', source_object=Object(type='ipv4', value='127.0.0.1', id='85803795-2026-4526-97bb-6221563bf05e'), destination_object=Object(type='report', value='87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7', id='9a94c8ad-7b85-41e4-bed5-13362c6a41ac'))
The fact will be replaced with the fact added to the platform and it will now have an id.
>>> print(f.id)
'5e533787-e71d-4ba4-9208-531f9baf8437'
A string representation of the fact will show a human readable version of the fact.
>>> str(f)
'(ipv4/127.0.0.1) -[seenIn/report]-> (report/87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7)'
You can specify origins, when creating facts:
>>> act.api.base.origin_map(c.config)
{'John Doe': '00000000-0000-0000-0000-000000000001', 'Test origin': '5da8b157-5129-4f2f-9b90-6d624d62eebe'}
>>> f = c.fact("mentions", origin=c.origin(name="Test origin")).source("report", "87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7").destination("ipv4", "127.0.0.1")
>>> f.serialize()
{'type': 'mentions', 'value': '', 'origin': '5da8b157-5129-4f2f-9b90-6d624d62eebe', 'accessMode': 'Public', 'sourceObject': {'type': 'report', 'value': '87428fc522803d31065e7bce3cf03fe475096631e5e07bbd7a0fde60c4cf25c7'}, 'destinationObject': {'type': 'ipv4', 'value': '127.0.0.1'}, 'bidirectionalBinding': False}
Use get() to get a fact by it's id.
>>> f = c.fact(id='5e533787-e71d-4ba4-9208-531f9baf8437').get()
Properties on objects can be retrieved by dot notation.
>>> f.type.name
'seenIn'
>>> f.value
'report'
Use meta() to create meta facts (facts about facts).
>>> f = c.fact(id='cf8a5e68-c87f-4595-ba19-cc85e01f2f13').get()
>>> import time
>>> meta = f.meta("observationTime", int(time.time()))
>>> meta
Fact(type='observationTime', value=1544785280, in_reference_to=ReferencedFact(type='seenIn', value='report', id='cf8a5e68-c87f-4595-ba19-cc85e01f2f13'))
As with facts, the meta fact is not sent to the backend, and you must use add() to submit it to the platform.
>>> meta.add()
Fact(type='observationTime', value='1544785280', in_reference_to=ReferencedFact(type='seenIn', value='report', id='cf8a5e68-c87f-4595-ba19-cc85e01f2f13'))
Use get_meta() to get meta facts (facts about facts).
>>> f = c.fact(id='6d80469f-bc73-4520-a82a-7667a6526362').get()
>>> meta = f.get_meta()
>>> print(meta[0])
[observationTime/2018-12-12T13:42:17.526912]
Use retract() to retract a fact.. The fact must have an id, either by specyfing it directly, or retriving the fact from a search.
>>> f = c.fact(id='1ba6c36a-8300-4ea1-aded-03ee80083dff')
>>> f.retract()
>>> objects = c.object_search(object_value="127.0.0.1", after="2016-09-28T21:26:22Z")
>>> len(objects)
1
>>> objects[0].type.name
'ipv4'
>>> objects[0].value
'127.0.0.1'
>>> objects[0].statistics[0].type.name
'DNSRecord'
>>> objects[0].statistics[0].count
131
>>> objects[0].statistics[1].type.name
'seenIn'
>>> objects[0].statistics[1].count
114
>>> object_type = c.object_type("fqdn").add()
Search fact and limit search by using the parameters.
>>> help(c.fact_search)
Help on method fact_search in module act.helpers:
fact_search(keywords='', object_type=[], fact_type=[], object_value=[], fact_value=[], organization=[], source=[], include_retracted=None, before=None, after=None, limit=None) method of act.helpers.Act instance
Search objects
Args:
keywords (str): Only return Facts matching a keyword query
object_type (str[] | str): Only return Facts with objects having a specific
ObjectType
fact_type (str[] | str): Only return Facts having a specific FactType
object_value (str[] | str): Only return Facts with objects matching a specific
value
fact_value (str[] | str): Only return Facts matching a specific value
organization (str[] | str): Only return Facts belonging to
a specific Organization
source (str[] | str): Only return Facts coming from a specific Source
include_retracted (bool): Include retracted Facts (default=False)
before (timestamp): Only return Facts added before a specific
timestamp. Timestamp is on this format:
2016-09-28T21:26:22Z
after (timestamp): Only return Facts added after a specific
timestamp. Timestamp is on this format:
2016-09-28T21:26:22Z
limit (integer): Limit the number of returned Objects
(default 25). Limit must be <= 10000.
All arguments are optional.
Returns ActResultSet of Facts.
By default the search will return and ActResultSet with 25 itmes.
>>> facts = c.fact_search(fact_type="seenIn", fact_value="report")
>>> len(facts)
25
>>> facts.size
25
>>> facts.count
820304
The complete property can be used to to check whether the result returned all available items.
>>> facts.complete
False
Use the limit parameter to get more items.
>>> facts = c.fact_search(fact_type="seenIn", fact_value="report", object_value="127.0.0.1", limit=2000)
>>> facts.size
119
>>> facts.complete
True
Get all object types.
>>> object_types = c.get_object_types()
>>> len(object_types)
46
>>> len(object_types)
46
>>> object_types[0].name
'technique'
>>> object_types[0].validator
'RegexValidator'
>>> object_types[0].validator_parameter
'(.|\\n)*'
The act platform has support for graph queries using the Gremlin Query language.
Use the traverse() function from an object to perform a graph query.
>>> path = c.object("ipv4", "127.0.0.220").traverse('g.bothE("seenIn").bothV().path().unfold()')
>>> type(path[0])
<class 'act.obj.Object'>
>>> type(path[1])
<class 'act.fact.Fact'>
You will normally want to use unfold() in the gremlin query to make sure you recive objects and facts.
Here is an example querying for threat actor aliases.
The graph of this will look like the screen shot below.
>>> aliases = c.object("threatActor", "APT 29").traverse('g.repeat(outE("threatActorAlias").outV()).until(cyclicPath()).path().unfold()')
>>> obj = [obj.value for obj in aliases if isinstance(obj, act.obj.Object)]
>>> obj
['APT 29', 'OfficeMonkeys', 'APT 29', 'APT 29', 'The Dukes', 'APT 29', 'APT 29', 'Hammer Toss', 'APT 29', 'APT 29', 'EuroAPT', 'APT 29', 'APT 29', 'CozyDuke', 'APT 29', 'APT 29', 'Office Monkeys', 'APT 29', 'APT 29', 'CozyCar', 'APT 29', 'APT 29', 'APT29', 'APT 29', 'APT 29', 'Dukes', 'APT 29', 'APT 29', 'Cozy Duke', 'APT 29', 'APT 29', 'Cozer', 'APT 29', 'APT 29', 'CozyBear', 'APT 29', 'APT 29', 'Cozy Bear', 'APT 29', 'APT 29', 'SeaDuke', 'APT 29', 'APT 29', 'Group 100', 'APT 29', 'APT 29', 'Minidionis', 'APT 29', 'APT 29', 'The Dukes', 'APT29', 'APT 29', 'APT 29', 'The Dukes', 'APT29', 'The Dukes', 'APT 29', 'CozyDuke', 'APT29', 'APT 29', 'APT 29', 'CozyDuke', 'APT29', 'CozyDuke', 'APT 29', 'APT29', 'The Dukes', 'APT29', 'APT 29', 'APT29', 'The Dukes', 'APT 29', 'APT 29', 'APT29', 'Cozy Bear', 'APT 29', 'APT 29', 'APT29', 'Cozy Bear', 'APT29', 'APT 29', 'APT29', 'CozyDuke', 'APT 29', 'APT 29', 'APT29', 'CozyDuke', 'APT29', 'APT 29', 'Cozy Bear', 'APT29', 'APT 29', 'APT 29', 'Cozy Bear', 'APT29', 'Cozy Bear', 'APT 29', 'The Dukes', 'APT29', 'Cozy Bear', 'APT 29', 'APT 29', 'The Dukes', 'APT29', 'Cozy Bear', 'APT29', 'APT 29', 'The Dukes', 'APT29', 'CozyDuke', 'APT 29', 'APT 29', 'The Dukes', 'APT29', 'CozyDuke', 'APT29', 'APT 29', 'CozyDuke', 'APT29', 'The Dukes', 'APT29', 'APT 29', 'CozyDuke', 'APT29', 'The Dukes', 'APT 29', 'APT 29', 'CozyDuke', 'APT29', 'Cozy Bear', 'APT 29', 'APT 29', 'CozyDuke', 'APT29', 'Cozy Bear', 'APT29', 'APT 29', 'Cozy Bear', 'APT29', 'The Dukes', 'APT29', 'APT 29', 'Cozy Bear', 'APT29', 'The Dukes', 'APT 29', 'APT 29', 'Cozy Bear', 'APT29', 'CozyDuke', 'APT 29', 'APT 29', 'Cozy Bear', 'APT29', 'CozyDuke', 'APT29']
>>> set(obj)
{'Office Monkeys', 'EuroAPT', 'Minidionis', 'APT29', 'OfficeMonkeys', 'Hammer Toss', 'CozyCar', 'The Dukes', 'Cozer', 'CozyBear', 'Cozy Bear', 'SeaDuke', 'Group 100', 'Dukes', 'CozyDuke', 'Cozy Duke', 'APT 29'}
Most instances are bootstrapped with a type system that will cover most use cases. However, it is also possible to extend the system with additional objectTypes / factTypes.
You can add objects by creating ObjectType object and executing add(). There is also a shortcut available on the client (object_type) which can used like this:
>>> c.object_type(name="filename", validator_parameter='.+').add()
ObjectType(name='filename', id='432c6d8a-542c-4374-94d1-b14e95139877', validator_parameter='.+', namespace=NameSpace(name='Global', id='00000000-0000-0000-0000-000000000000'))
The validator_parameter specifies what values that are allowed on this object. In this example, any non-empty values are allowed.
Facts specifies relation to one or two Objects and to add facts there must be FactTypes that specifies these bindings. There is a helper function that will create a Fact Type with bindings between all exisiting object types in the system:
>>>> c.create_fact_type_all_bindings("droppedBy", '.*')
(...)
However, on production systems it is advisable to only create bindings between objects that makes sense for the given Fact Type, like this:
>>> object_bindings = [{
"destinationObjectType": "hash",
"sourceObjectType": "filename"
}]
>>> c.create_fact_type("droppedBy", '.*', object_bindings = object_bindings)
FactType(name='droppedBy', id='cbc49137-3c52-4655-8b47-386d31de231a', validator_parameter='.*', relevant_object_bindings=[RelevantObjectBindings(source_object_type='432c6d8a-542c-4374-94d1-b14e95139877', destination_object_type='e4b673b6-7a59-4fca-b8eb-ff4489501cf5')], namespace=NameSpace(name='Global', id='00000000-0000-0000-0000-000000000000'))
The bindings will be created using a combination of all source/destination objects for each entry.
It is also possible to specify bidirectional bindings like this:
>>> object_bindings = [{
"bidirectional": true,
"destinationObjectType": "threatActor",
"sourceObjectType": "threatActor"
}]
Facts are immutable, so it is not possible to update the ObjectType and FactType validators, as this might lead to an incositence state. However, it is possible to add Object Bindings to existing facts. This function require the objets to be retrived first:
>>> dropped_by = [ft for ft in c.get_fact_types() if ft.name == "droppedBy"][0]
>>> hash = [ot for ot in c.get_object_types() if ot.name == "hash"][0]
>>> filename = [ot for ot in c.get_object_types() if ot.name == "filename"][0]
>>> dropped_by.id
'18b0f70e-82dc-4904-b745-d20b0ac54adf'
>>>> dropped_by.add_binding(source_object_type=filename, destination_object_type=hash)
The platforms supports origin to support where the fact originates from. If now origin is given when creating a fact, the origin will be the user itself.
You can list origins using get_origins():
>>> c.get_origins()
[Origin(name='John Doe', id='00000000-0000-0000-0000-000000000001', namespace=NameSpace(name='Global', id='00000000-0000-0000-0000-000000000000'), organization=Organization(name='Test Organization 1', id='00000000-0000-0000-0000-000000000001'), trust=0.8)]
>>> o = c.origin("Test origin", trust=0.5, description="My test origin")
>>> o.add()
Origin(name='Test origin', id='5da8b157-5129-4f2f-9b90-6d624d62eebe', namespace=NameSpace(name='Global', id='00000000-0000-0000-0000-000000000000'), organization=Organization(), description='My test origin', trust=0.5)
>>> o = c.origin(id="5da8b157-5129-4f2f-9b90-6d624d62eebe")
>>> o.get()
Origin(name='Test origin', id='5da8b157-5129-4f2f-9b90-6d624d62eebe', namespace=NameSpace(name='Global', id='00000000-0000-0000-0000-000000000000'), organization=Organization(), description='My test origin', trust=0.5)
Fact chain is currently a experimental concept, that supports chains of fact, where some of the objects in the chain can unknowns / placeholders.
The unknowns are marked using value "*". And after the chain is created they will get a special value "[placeholder[]]", where the HASH is caclulated based on the incoming/outgoing paths from the placeholder.
>>> facts = (
c.fact("observedIn").source("uri", "http://uri.no").destination("incident", "*"),
c.fact("targets").source("incident", "*").destination("organization", "*"),
c.fact("memberOf").source("organization", "*").destination("sector", "energy"),
)
>>> chain = act.fact.fact_chain(*facts)
>>> for fact in chain:
fact.add()
This feature should be considered experimental and are subject to change. It is implemented client side and the backend does not have the notion of what a fact chain is at the moment, but the frontned will currently show the value in a more user friendly way.
Also note that adding facts in a chain as shown above is NOT atomic, and might lead to inconsistencies if some of the facts does not pass validation in the backend.
Tests (written in pytest) are contained in the test/ folder. Mock objects are available for most API requests in the test/data/ folder.
This command will execute the tests using both python2 and python3 (requires pytest, python2 and python3).
test/run-tests.sh