Integration Details

ThreatQuotient provides the following details for this integration:

Current Integration Version	1.0.0
Compatible with ThreatQ Versions	>= 5.12.1
Support Tier	ThreatQ Supported

Introduction

The MITRE Adversarial Threat Landscape for AI Systems (ATLAS) CDF integration allows teams to ingest the MITRE ATLAS knowledgebase of artificial intelligence techniques used by threat actors. These techniques are mapped alongside the MITRE ATT&CK framework, providing a view of the tactics, techniques, and procedures (TTPs) used by adversaries to target AI systems.

MITRE ATLAS is a comprehensive knowledge base designed to help protect artificial intelligence systems by cataloging real-world attacks on AI, providing insights into adversary tactics and techniques.

The integration provides the following feeds:

• MITRE ATLAS - ingests techniques used by threat actors targeting AI systems.
• MITRE ATLAS Reports - ingests the case studies/reports generated by OpenCTI that are linked to the MITRE ATLAS framework.

The integration ingests the following object types:

• Attack Patterns
• Courses of Action
• Incidents
• Identities 
• Reports
• Vulnerabilities

Prerequisites

The following is not required but recommended for the integration:

• GitHub API Token - In order to pull data from the GitHub API with the MITRE ATLAS Reports feed, you may need to authenticate with a GitHub API Token.  See the Known Issues / Limitations section for more details. 

Installation

Perform the following steps to install the integration:

The same steps can be used to upgrade the integration to a new version.

1. Log into https://marketplace.threatq.com/.
2. Locate and download the integration yaml file.
3. Navigate to the integrations management page on your ThreatQ instance.
4. Click on the Add New Integration button.
5. Upload the integration yaml file using one of the following methods:
   • Drag and drop the file into the dialog box
   • Select Click to Browse to locate the file on your local machine
6. Select the individual feeds to install, when prompted, and click Install.

   ThreatQ will inform you if the feed already exists on the platform and will require user confirmation before proceeding. ThreatQ will also inform you if the new version of the feed contains changes to the user configuration. The new user configurations will overwrite the existing ones for the feed and will require user confirmation before proceeding.

The feed(s) will be added to the integrations page. You will still need to configure and then enable the feed.

Configuration

ThreatQuotient does not issue API keys for third-party vendors. Contact the specific vendor to obtain API keys and other integration-related credentials.

To configure the integration:

1. Navigate to your integrations management page in ThreatQ.
2. Select the OSINT option from the Category dropdown (optional).

   If you are installing the integration for the first time, it will be located under the Disabled tab.

3. Click on the integration entry to open its details page.
4. Enter the following parameters under the Configuration tab:

   MITRE ATLAS Parameters

   Parameter	Description
   Save Intrusion Set As	Select how to ingest Intrusion Sets into the ThreatQ platform. Options include: Adversaries Intrusion Sets ThreatQuotient recommends users select the Adversaries option to create a consolidated profile.
   Disable Proxies	Enable this option if the feed should not honor proxies set in the ThreatQ UI.
   Enable SSL Certificate  Verification	Enable this for the feed to validate the host-provided SSL certificate.

   
   

   MITRE ATLAS Reports Parameters

   Parameter	Description
   GitHub API Token	Optional and Recommended - Enter your GitHub API Token (Personal Access Token). The MITRE ATLAS Reports feed will fetch the MITRE ATLAS Report data from their public GitHub repository. These reports are generated by OpenCTI and contain information linked to the MITRE ATLAS framework. The ATLAS Reports reside within a GitHub repository and are updated on a semi-regular basis. In order to pull data from the GitHub API, you may need to authenticate with a GitHub API Token. This will increase the rate limit from 60 requests per hour to 5000 requests per hour. If you are experiencing rate limiting issues, please consider adding your GitHub API Token to the feed configuration. Authentication is not required to access the MITRE ATLAS data, but is highly recommended. If you are experiencing rate limiting issues, you can add a GitHub API Token to the feed configuration. Unauthenticated users have a rate limit of 60 requests per hour, while authenticated users have a rate limit of 5000 requests per hour.
   Save Intrusion Set As	Select how to ingest Intrusion Sets into the ThreatQ platform. Options include: Adversaries Intrusion Sets ThreatQuotient recommends users select the Adversaries option to create a consolidated profile.
   Disable Proxies	Enable this option if the feed should not honor proxies set in the ThreatQ UI.
   Enable SSL Certificate  Verification	Enable this for the feed to validate the host-provided SSL certificate.

   
   
5. Review any additional settings, make any changes if needed, and click on Save.
6. Click on the toggle switch, located above the Additional Information section, to enable it.

ThreatQ Mapping

MITRE ATLAS

The MTIRE ATLAS feed ingests techniques used by threat actors targeting AI systems. These techniques will be ingested into ThreatQ as Attack Pattern objects, and will include context such as tactics, references, and descriptions.

GET https://raw.githubusercontent.com/mitre-atlas/atlas-navigator-data/main/dist/stix-atlas.json

Sample Response:

{
    "type": "bundle",
    "id": "bundle--a9061128-f24c-414d-ac6b-4dc6a3329e3b",
    "objects": [
        {
            "type": "x-mitre-tactic",
            "spec_version": "2.1",
            "id": "x-mitre-tactic--4bf0ef0b-3de4-46fe-9a0d-3b1dea352a30",
            "created": "2024-06-24T20:23:47.456097Z",
            "modified": "2024-06-24T20:23:47.456097Z",
            "name": "Reconnaissance",
            "description": "The adversary is trying to gather information about the machine learning system they can use to plan future operations.\n\nReconnaissance consists of techniques that involve adversaries actively or passively gathering information that can be used to support targeting.\nSuch information may include details of the victim organizations' machine learning capabilities and research efforts.\nThis information can be leveraged by the adversary to aid in other phases of the adversary lifecycle, such as using gathered information to obtain relevant ML artifacts, targeting ML capabilities used by the victim, tailoring attacks to the particular models used by the victim, or to drive and lead further Reconnaissance efforts.\n",
            "external_references": [
                {
                    "source_name": "mitre-atlas",
                    "url": "https://atlas.mitre.org/tactics/AML.TA0002",
                    "external_id": "AML.TA0002"
                }
            ],
            "x_mitre_shortname": "reconnaissance"
        },
        {
            "type": "x-mitre-tactic",
            "spec_version": "2.1",
            "id": "x-mitre-tactic--6b232c1e-ada7-4cd4-b538-7a1ef6193e2f",
            "created": "2024-06-24T20:23:47.45631Z",
            "modified": "2024-06-24T20:23:47.45631Z",
            "name": "Resource Development",
            "description": "The adversary is trying to establish resources they can use to support operations.\n\nResource Development consists of techniques that involve adversaries creating,\npurchasing, or compromising/stealing resources that can be used to support targeting.\nSuch resources include machine learning artifacts, infrastructure, accounts, or capabilities.\nThese resources can be leveraged by the adversary to aid in other phases of the adversary lifecycle, such as [ML Attack Staging](/tactics/AML.TA0001).\n",
            "external_references": [
                {
                    "source_name": "mitre-atlas",
                    "url": "https://atlas.mitre.org/tactics/AML.TA0003",
                    "external_id": "AML.TA0003"
                }
            ],
            "x_mitre_shortname": "resource-development"
        }
    ]
}

The data is mapped via the default STIX 2.1 mapping, which can be found on the ThreatQ Help Center: https://helpcenter.threatq.com/ThreatQ_Platform/System_Objects/STIX/STIX_2.1.htm.

MITRE ATLAS Reports

The MITRE ATLAS Reports feed ingests the case studies/reports generated by OpenCTI that are linked to the MITRE ATLAS framework. These reports will be ingested into ThreatQ as Report objects, and will include context such as the report title, description, and references.

The feed makes requests to the following GitHub API Endpoints:

GET https://api.github.com/rate_limit - fetches the rate limits for the current user (or no user).
GET https://api.github.com/repos/mitre-atlas/atlas-navigator-data/commits - fetches the commits for the repository's "dist/opencti-bundles" content directory.
GET https://api.github.com/repos/mitre-atlas/atlas-navigator-data/commits/{{ sha_hash }} - fetches the changed files for each commit.
GET https://raw.githubusercontent.com/mitre-atlas/atlas-navigator-data/main/dist/opencti-bundles/{{ file_name }} - fetches the raw JSON file for each case study.

Sample Response:

{
    "type": "bundle",
    "id": "bundle--ce0e48fe-3174-4e6d-9887-d87f2b286f63",
    "objects": [
        {
            "id": "report--2a5fab52-5c12-56b2-8ca2-5e1e3c8ae6bb",
            "spec_version": "2.1",
            "revoked": false,
            "x_opencti_reliability": "A - Completely reliable",
            "confidence": 75,
            "created": "2020-01-12T12:00:00.000Z",
            "modified": "2024-05-22T17:24:46.467Z",
            "name": "MITRE ATLAS Case Study: Microsoft Azure Service Disruption",
            "description": "The Microsoft AI Red Team performed a red team exercise on an internal Azure service with the intention of disrupting its service. This operation had a combination of traditional ATT&CK enterprise techniques such as finding valid account, and exfiltrating data -- all interleaved with adversarial ML specific steps such as offline and online evasion examples.",
            "report_types": [
                "internal-report"
            ],
            "published": "2020-01-12T12:00:00.000Z",
            "x_opencti_workflow_id": "4c154301-4863-42d1-ba1a-bcf5cff32385",
            "labels": [
                "mitre atlas source",
                "aml.cs0010"
            ],
            "external_references": [
                {
                    "source_name": "AML.CS0010",
                    "url": "https://atlas.mitre.org/studies/AML.CS0010/"
                }
            ],
            "x_opencti_id": "4f8938d4-2da4-445b-8770-5f7f18cd9248",
            "x_opencti_type": "Report",
            "type": "report",
            "created_by_ref": "identity--48b4f20a-ddd3-5524-a646-55dd68b62ede",
            "object_marking_refs": [
                "marking-definition--613f2e26-407d-48c7-9eca-b8e91df99dc9"
            ],
            "object_refs": [
                "attack-pattern--18c33065-849f-5705-b4d4-28a08470fba6",
                "attack-pattern--2dc305cb-dd66-55b6-a0c8-87dbf69bae66"
                "incident--8458a7d3-dc00-5460-9d32-8db303ab7415"
            ]
        },
        {
            "id": "attack-pattern--18c33065-849f-5705-b4d4-28a08470fba6",
            "spec_version": "2.1",
            "revoked": false,
            "confidence": 0,
            "created": "2023-10-31T13:48:07.631Z",
            "modified": "2023-11-20T18:19:47.016Z",
            "name": "Exfiltration via Cyber Means",
            "description": "Adversaries may exfiltrate ML artifacts or other information relevant to their goals via traditional cyber means.\n\nSee the ATT&CK [Exfiltration](https://attack.mitre.org/tactics/TA0010/) tactic for more information.",
            "x_mitre_id": "AML.T0025",
            "labels": [
                "atlas"
            ],
            "kill_chain_phases": [
                {
                    "kill_chain_name": "mitre-atlas",
                    "phase_name": "exfiltration",
                    "x_opencti_order": 13
                }
            ],
            "external_references": [
                {
                    "source_name": "mitre-atlas",
                    "url": "https://atlas.mitre.org/techniques/AML.T0025",
                    "external_id": "AML.T0025"
                }
            ],
            "x_opencti_id": "ec43ad24-eebc-4442-b1cd-7d82123b9313",
            "x_opencti_type": "Attack-Pattern",
            "type": "attack-pattern",
            "object_marking_refs": [
                "marking-definition--613f2e26-407d-48c7-9eca-b8e91df99dc9",
                "marking-definition--c4099854-e00c-5dbc-b0fe-4c9909920101"
            ]
        },
        {
            "id": "attack-pattern--2dc305cb-dd66-55b6-a0c8-87dbf69bae66",
            "spec_version": "2.1",
            "revoked": false,
            "confidence": 0,
            "created": "2023-10-31T13:48:07.628Z",
            "modified": "2023-11-20T18:19:43.967Z",
            "name": "Valid Accounts",
            "description": "Adversaries may obtain and abuse credentials of existing accounts as a means of gaining Initial Access.\nCredentials may take the form of usernames and passwords of individual user accounts or API keys that provide access to various ML resources and services.\n\nCompromised credentials may provide access to additional ML artifacts and allow the adversary to perform [Discover ML Artifacts](/techniques/AML.T0007).\nCompromised credentials may also grant and adversary increased privileges such as write access to ML artifacts used during development or production.",
            "x_mitre_id": "AML.T0012",
            "labels": [
                "atlas"
            ],
            "kill_chain_phases": [
                {
                    "kill_chain_name": "mitre-atlas",
                    "phase_name": "initial-access",
                    "x_opencti_order": 3
                }
            ],
            "external_references": [
                {
                    "source_name": "mitre-atlas",
                    "url": "https://atlas.mitre.org/techniques/AML.T0012",
                    "external_id": "AML.T0012"
                }
            ],
            "x_opencti_id": "eb0a6b17-ae24-4d15-be8f-c70ade49345d",
            "x_opencti_type": "Attack-Pattern",
            "type": "attack-pattern",
            "object_marking_refs": [
                "marking-definition--613f2e26-407d-48c7-9eca-b8e91df99dc9",
                "marking-definition--c4099854-e00c-5dbc-b0fe-4c9909920101"
            ]
        },
        {
            "id": "incident--8458a7d3-dc00-5460-9d32-8db303ab7415",
            "spec_version": "2.1",
            "revoked": false,
            "confidence": 75,
            "created": "2024-02-06T15:29:37.168Z",
            "modified": "2024-02-06T16:44:38.985Z",
            "name": "2020 Microsoft Azure Service Disruption",
            "description": "The Microsoft AI Red Team performed a red team exercise on an internal Azure service with the intention of disrupting its service. This operation had a combination of traditional ATT&CK enterprise techniques such as finding valid account, and exfiltrating data -- all interleaved with adversarial ML specific steps such as offline and online evasion examples.",
            "first_seen": "2020-01-12T12:00:00.000Z",
            "incident_type": "research-finding",
            "labels": [
                "aml.cs0010"
            ],
            "external_references": [
                {
                    "source_name": "AML.CS0010",
                    "url": "https://atlas.mitre.org/studies/AML.CS0010/"
                }
            ],
            "x_opencti_id": "7f59d9c4-f27e-4553-91be-788af57929c7",
            "x_opencti_type": "Incident",
            "type": "incident",
            "created_by_ref": "identity--48b4f20a-ddd3-5524-a646-55dd68b62ede",
            "object_marking_refs": [
                "marking-definition--613f2e26-407d-48c7-9eca-b8e91df99dc9"
            ]
        }
    ]
}

The data is mapped via the default STIX 2.1 mapping, which can be found on the ThreatQ Help Center: https://helpcenter.threatq.com/ThreatQ_Platform/System_Objects/STIX/STIX_2.1.htm.

Average Feed Run

Object counts and Feed runtime are supplied as generalities only - objects returned by a provider can differ based on credential configurations and Feed runtime may vary based on system resources and load.

MITRE ATLAS

Metric	Result
Run Time	1 minute
Attack Patterns	82
Attack Pattern Attributes	428
Courses of Action	20
Course of Action Attributes	60

MITRE ATLAS Reports

Metric	Result
Run Time	1 minute
Attack Patterns	32
Attack Pattern Attributes	234
Identities	2
Incidents	7
Incident Attributes	30
Reports	8
Report Attributes	43
Vulnerabilities	2
Vulnerability Attributes	8

Known Issues / Limitations

• MITRE ATLAS Reports - The ATLAS Reports reside within a GitHub repository and are updated on a semi-regular basis. In order to pull data from the GitHub API, you may need to authenticate with a GitHub API Token. This will increase the rate limit from 60 requests per hour to 5000 requests per hour. If you are experiencing rate limiting issues, please consider adding your GitHub API Token to the feed configuration.
  While authentication is not required, it is highly recommended when accessing the MITRE ATLAS data. If you are experiencing rate limiting issues, you can add a GitHub API Token to the feed configuration. Unauthenticated users have a rate limit of 60 requests per hour, while authenticated users have a rate limit of 5000 requests per hour.

Change Log

• Version 1.0.0
  • Initial release