Technique library
How AI systems get attacked
Every story, CVE and tool on ThreatDeck is mapped to the public frameworks defenders already use. Pick a technique to see what is happening to it right now.
OWASP Top 10 for LLM Applications (2025)
OWASP- LLM01 Prompt Injection 10 stories 15 tools
- LLM02 Sensitive Information Disclosure 5 stories 5 tools
- LLM03 Supply Chain 8 stories 7 tools
- LLM04 Data and Model Poisoning 0 stories 0 tools
- LLM05 Improper Output Handling 0 stories 3 tools
- LLM06 Excessive Agency 1 story 2 tools
- LLM07 System Prompt Leakage 1 story 0 tools
- LLM08 Vector and Embedding Weaknesses 0 stories 0 tools
- LLM09 Misinformation 1 story 3 tools
- LLM10 Unbounded Consumption 0 stories 0 tools
OWASP Top 10 for Agentic Applications (2026)
OWASP- ASI01 Agent Goal Hijack 2 stories 4 tools
- ASI02 Tool Misuse and Exploitation 10 stories 4 tools
- ASI03 Identity and Privilege Abuse 3 stories 1 tool
- ASI04 Agentic Supply Chain Vulnerabilities 2 stories 7 tools
- ASI05 Unexpected Code Execution 3 stories 2 tools
- ASI06 Memory and Context Poisoning 2 stories 0 tools
- ASI07 Insecure Inter-Agent Communication 1 story 0 tools
- ASI08 Cascading Failures 0 stories 0 tools
- ASI09 Human-Agent Trust Exploitation 0 stories 0 tools
- ASI10 Rogue Agents 9 stories 0 tools
MITRE ATLAS
atlas.mitre.orgIn the news or covered by tracked tools
- AML.T0010 Supply Chain Compromise 3 stories 3 tools
- AML.T0020 Training Data Poisoning 0 stories 1 tool
- AML.T0024 Exfiltration via AI Inference API 2 stories 0 tools
- AML.T0034 Cost Harvesting 0 stories 1 tool
- AML.T0043 Craft Adversarial Data 1 story 2 tools
- AML.T0051 LLM Prompt Injection 3 stories 3 tools
- AML.T0054 LLM Jailbreak 2 stories 4 tools
- AML.T0080 AI Agent Context Poisoning 1 story 0 tools
- AML.T0083 Credentials from AI Agent Configuration 2 stories 1 tool
- AML.T0086 Exfiltration via AI Agent Tool Invocation 2 stories 0 tools
- AML.T0098 AI Agent Tool Credential Harvesting 1 story 0 tools
- AML.T0103 Deploy AI Agent 5 stories 0 tools
- AML.T0110 AI Agent Tool Poisoning 1 story 4 tools
All other ATLAS techniques (116)
AI Attack Adaptation
- AML.T0005 Replicate AI Model Behavior 0 stories 0 tools
- AML.T0018 Manipulate AI Model 0 stories 0 tools
- AML.T0042 Verify Attack 0 stories 0 tools
- AML.T0065 LLM Prompt Crafting 0 stories 0 tools
- AML.T0066 Retrieval Content Crafting 0 stories 0 tools
- AML.T0088 Generate Deepfakes 0 stories 0 tools
- AML.T0102 Generate Malicious Commands 0 stories 0 tools
- AML.T0117 Autonomous Attack-Path Adaptation 0 stories 0 tools
- AML.T0118 Autonomous AI Agent Communication 0 stories 0 tools
- AML.T0124 Autonomous Attack Orchestration 0 stories 0 tools
AI Model Access
- AML.T0040 AI Model Inference API Access 0 stories 0 tools
- AML.T0041 Physical Environment Access 0 stories 0 tools
- AML.T0044 Full AI Model Access 0 stories 0 tools
- AML.T0047 AI-Enabled Product or Service 0 stories 0 tools
Collection
- AML.T0035 AI Artifact Collection 0 stories 0 tools
- AML.T0036 Data from Information Repositories 0 stories 0 tools
- AML.T0037 Data from Local System 0 stories 0 tools
- AML.T0085 Data from AI Services 0 stories 0 tools
- AML.T0126 Automated Collection 0 stories 0 tools
- AML.T0127 Data Staged 0 stories 0 tools
- AML.T0135 Outputs from AI Models 0 stories 0 tools
- AML.T0137 Training Data Information from AI Models 0 stories 0 tools
Command and Control
- AML.T0072 Cyber Communication Channel 0 stories 0 tools
- AML.T0096 AI Service API 0 stories 0 tools
- AML.T0108 AI Agent 0 stories 0 tools
- AML.T0114 AI Service Web Interface 0 stories 0 tools
- AML.T0120 AI Artifact Repository 0 stories 0 tools
Credential Access
- AML.T0055 Unsecured Credentials 0 stories 0 tools
- AML.T0082 RAG Credential Harvesting 0 stories 0 tools
- AML.T0090 OS Credential Dumping 0 stories 0 tools
- AML.T0106 Exploitation for Credential Access 0 stories 0 tools
- AML.T0113 Steal Web Session Cookie 0 stories 0 tools
Defense Evasion
- AML.T0067 LLM Trusted Output Components Manipulation 0 stories 0 tools
- AML.T0068 LLM Prompt Obfuscation 0 stories 0 tools
- AML.T0071 False RAG Entry Injection 0 stories 0 tools
- AML.T0073 Impersonation 0 stories 0 tools
- AML.T0074 Masquerading 0 stories 0 tools
- AML.T0076 Corrupt AI Model 0 stories 0 tools
- AML.T0092 Manipulate User LLM Chat History 0 stories 0 tools
- AML.T0094 Delay Execution of LLM Instructions 0 stories 0 tools
- AML.T0097 Virtualization/Sandbox Evasion 0 stories 0 tools
- AML.T0107 Exploitation for Defense Evasion 0 stories 0 tools
- AML.T0109 AI Supply Chain Rug Pull 0 stories 0 tools
- AML.T0111 AI Supply Chain Reputation Inflation 0 stories 0 tools
- AML.T0123 Obfuscated Files or Information 0 stories 0 tools
- AML.T0129 Triggers in Multimodal Inputs 0 stories 0 tools
- AML.T0134 AI Targeted Cloaking 0 stories 0 tools
- AML.T0138 Indicator Removal 0 stories 0 tools
- AML.T0140 Centralized AI Service Request Routing 0 stories 0 tools
- AML.T0141 AI Service Request Metadata Sanitization 0 stories 0 tools
Discovery
- AML.T0007 Discover AI Artifacts 0 stories 0 tools
- AML.T0013 Discover AI Model Ontology 0 stories 0 tools
- AML.T0014 Discover AI Model Family 0 stories 0 tools
- AML.T0062 Discover LLM Hallucinations 0 stories 0 tools
- AML.T0063 Discover AI Model Outputs 0 stories 0 tools
- AML.T0069 Discover LLM System Information 0 stories 0 tools
- AML.T0075 Enterprise Resource Discovery 0 stories 0 tools
- AML.T0084 Discover AI Agent Configuration 0 stories 0 tools
- AML.T0089 Enterprise Environment Discovery 0 stories 0 tools
- AML.T0133 Discover AI Agent Runtime Capabilities 0 stories 0 tools
- AML.T0144 Recover AI Model Internals 0 stories 0 tools
Execution
- AML.T0011 User Execution 0 stories 0 tools
- AML.T0050 Command and Scripting Interpreter 0 stories 0 tools
- AML.T0053 AI Agent Tool Invocation 0 stories 0 tools
- AML.T0100 AI Agent Clickbait 0 stories 0 tools
Exfiltration
- AML.T0025 Exfiltration via Cyber Means 0 stories 0 tools
- AML.T0077 LLM Response Rendering 0 stories 0 tools
- AML.T0136 Exfiltration via AI User Interface 0 stories 0 tools
Impact
- AML.T0029 Denial of AI Service 0 stories 0 tools
- AML.T0031 Erode AI Model Integrity 0 stories 0 tools
- AML.T0046 Spamming AI System with Chaff Data 0 stories 0 tools
- AML.T0048 External Harms 0 stories 0 tools
- AML.T0059 Erode Dataset Integrity 0 stories 0 tools
- AML.T0101 Data Destruction via AI Agent Tool Invocation 0 stories 0 tools
- AML.T0112 Machine Compromise 0 stories 0 tools
- AML.T0130 AI Agent Response Biasing 0 stories 0 tools
Initial Access
- AML.T0012 Valid Accounts 0 stories 0 tools
- AML.T0015 Evade AI Model 0 stories 0 tools
- AML.T0049 Exploit Public-Facing Application 0 stories 0 tools
- AML.T0052 Phishing 0 stories 0 tools
- AML.T0078 Drive-by Compromise 0 stories 0 tools
- AML.T0093 Prompt Infiltration via Public-Facing Application 0 stories 0 tools
- AML.T0119 Exploit Automated Artifact Processing Pipeline 0 stories 0 tools
- AML.T0131 Crafted AI Assistant Links 0 stories 0 tools
- AML.T0132 Misconfigured or Publicly Exposed AI Services 0 stories 0 tools
Lateral Movement
- AML.T0091 Use Alternate Authentication Material 0 stories 0 tools
- AML.T0122 Exploitation of Remote Services 0 stories 0 tools
- AML.T0143 LLM Prompt Propagation 0 stories 0 tools
Persistence
- AML.T0061 LLM Prompt Self-Replication 0 stories 0 tools
- AML.T0070 RAG Poisoning 0 stories 0 tools
- AML.T0081 Modify AI Agent Configuration 0 stories 0 tools
- AML.T0099 AI Agent Tool Data Poisoning 0 stories 0 tools
- AML.T0121 AI Agent Environment Reconstruction 0 stories 0 tools
- AML.T0125 Create Account 0 stories 0 tools
- AML.T0142 Event Triggered Execution 0 stories 0 tools
- AML.T0145 Poison AI Agent Learning Environment 0 stories 0 tools
Privilege Escalation
- AML.T0105 Escape to Host 0 stories 0 tools
Reconnaissance
- AML.T0000 Search Open Technical Databases 0 stories 0 tools
- AML.T0001 Search Open AI Vulnerability Analysis 0 stories 0 tools
- AML.T0003 Search Victim-Owned Websites 0 stories 0 tools
- AML.T0004 Search Application Repositories 0 stories 0 tools
- AML.T0006 Active Scanning 0 stories 0 tools
- AML.T0064 Gather RAG-Indexed Targets 0 stories 0 tools
- AML.T0087 Gather Victim Identity Information 0 stories 0 tools
- AML.T0095 Search Open Websites/Domains 0 stories 0 tools
- AML.T0116 Autonomous Reconnaissance 0 stories 0 tools
Resource Development
- AML.T0002 Acquire Public AI Artifacts 0 stories 0 tools
- AML.T0008 Acquire Infrastructure 0 stories 0 tools
- AML.T0016 Obtain Capabilities 0 stories 0 tools
- AML.T0017 Develop Capabilities 0 stories 0 tools
- AML.T0021 Establish Accounts 0 stories 0 tools
- AML.T0060 Publish Hallucinated Entities 0 stories 0 tools
- AML.T0079 Stage Capabilities 0 stories 0 tools
- AML.T0115 Publish Poisoned AI Artifacts 0 stories 0 tools
- AML.T0128 Compromise Infrastructure 0 stories 0 tools
- AML.T0139 Compromise Accounts 0 stories 0 tools