---
reference_key: mitre-atlas
title: "MITRE ATLAS --- Adversarial ML"
url: "https://atlas.mitre.org/"
availability: available
capture_method: github_yaml_mirror
captured_at: 2026-06-23T03:30:32Z
document_id: REF-mitre-atlas
---

# MITRE ATLAS --- Adversarial ML

| Field | Value |
|-------|-------|
| **Reference key** | `mitre-atlas` |
| **Availability** | AVAILABLE |
| **Capture method** | `github_yaml_mirror` |
| **Source type** | remote_url |
| **URL / path** | https://atlas.mitre.org/ |
| **Captured (UTC)** | 2026-06-23T03:30:32Z |
| **Content type** | text/plain; charset=utf-8 |
| **HTTP status** | 200 |

## Boardroom Citation Context

Adversarial ML threat mapping; P2 red-team scenarios.

## Source Location

https://atlas.mitre.org/


## Capture Provenance

| Field | Value |
|-------|-------|
| **Capture method** | `github_yaml_mirror` |
| **Primary URL** | https://atlas.mitre.org/ |
| **Primary result** | partial |
| **Fallback URL** | https://raw.githubusercontent.com/mitre-atlas/atlas-data/main/dist/v6/ATLAS-2026.05.yaml |

Primary https://atlas.mitre.org/ is a JavaScript SPA with no static HTML body. MITRE publishes canonical ATLAS matrix data in the public mitre-atlas/atlas-data repository.

**Primary failure:** HTML retrieved but yielded minimal extractable text (JavaScript-rendered or paywall shell).




## Captured Content

```text
MITRE ATLAS canonical data export (YAML mirror from mitre-atlas/atlas-data).
Description: Adversarial Threat Landscape for AI Systems
Version: 2026.05
Modified: 2026-05-27
Tactics indexed: 16
Techniques indexed: 101

Sample tactics:
- AI Model Access
- AI Attack Staging
- Reconnaissance
- Resource Development
- Initial Access
- Execution
- Persistence
- Defense Evasion
- Discovery
- Collection
- Exfiltration
- Impact
- Privilege Escalation
- Credential Access
- Command and Control

Sample techniques:
- Search Open Technical Databases
- Search Open AI Vulnerability Analysis
- Acquire Public AI Artifacts
- Search Victim-Owned Websites
- Search Application Repositories
- Create Proxy AI Model
- Active Scanning
- Discover AI Artifacts
- Acquire Infrastructure
- AI Supply Chain Compromise
- User Execution
- Valid Accounts
- Discover AI Model Ontology
- Discover AI Model Family
- Evade AI Model
- Obtain Capabilities
- Develop Capabilities
- Manipulate AI Model
- Publish Poisoned Datasets
- Poison Training Data
- Establish Accounts
- Exfiltration via AI Inference API
- Exfiltration via Cyber Means
- Denial of AI Service
- Erode AI Model Integrity
```

---

*Archived reference article for AICSR-STUDY-2026-001 footnote corpus.*