There are 15 attacks in the cross-platform category. Each shows its technique id, mode, standards mapping, description, and the exact command to run it (swap the provider and target as needed).
​
CI/CD Injection
critical
ci-cd-injection · single-turn · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Injects malicious steps into CI/CD pipelines via AI code review or automated PR assistanceRun it:
ai-blackteam run -p anthropic -a ci-cd-injection -t "your target prompt"
​
Cloud Resource Abuse
critical
cloud-resource-abuse · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Abuses cloud APIs through AI to provision unauthorized resources, exfiltrate data, or escalate IAM privilegesRun it:
ai-blackteam run -p anthropic -a cloud-resource-abuse -t "your target prompt"
​
Container Escape
critical
container-escape · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Attempts container escape through AI tools by exploiting privileged mounts or kernel interfacesRun it:
ai-blackteam run -p anthropic -a container-escape -t "your target prompt"
​
Database Manipulation
critical
database-manipulation · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068SQL manipulation through AI database access tools to read, modify, or destroy dataRun it:
ai-blackteam run -p anthropic -a database-manipulation -t "your target prompt"
​
File System Traversal
critical
file-system-traversal · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Path traversal through AI file access tools to read or write files outside the intended directoryRun it:
ai-blackteam run -p anthropic -a file-system-traversal -t "your target prompt"
​
OAuth Manipulation
critical
oauth-manipulation · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Exploits OAuth flows in AI integrations to gain unauthorized access or escalate privilegesRun it:
ai-blackteam run -p anthropic -a oauth-manipulation -t "your target prompt"
​
Payment Manipulation
critical
payment-manipulation · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Manipulates payment APIs via AI to issue fraudulent refunds, transfer funds, or modify pricingRun it:
ai-blackteam run -p anthropic -a payment-manipulation -t "your target prompt"
​
API Abuse
high
api-abuse · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Exploits AI-connected APIs to trigger unintended actions beyond intended scopeRun it:
ai-blackteam run -p anthropic -a api-abuse -t "your target prompt"
​
Browser Exploitation
high
browser-exploitation · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Exploits AI browser or web automation tools to steal cookies, execute XSS, or perform unauthorized web actionsRun it:
ai-blackteam run -p anthropic -a browser-exploitation -t "your target prompt"
​
DNS Manipulation
high
dns-manipulation · single-turn · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068DNS rebinding or manipulation via AI tools to bypass network controls or exfiltrate dataRun it:
ai-blackteam run -p anthropic -a dns-manipulation -t "your target prompt"
​
Email Injection
high
email-injection · single-turn · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Exploits email-connected AI assistants to exfiltrate data or send unauthorized emailsRun it:
ai-blackteam run -p anthropic -a email-injection -t "your target prompt"
​
Slack Injection
high
slack-injection · single-turn · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Injects malicious instructions via Slack or Teams messages to hijack AI assistant actionsRun it:
ai-blackteam run -p anthropic -a slack-injection -t "your target prompt"
​
Webhook Exploitation
high
webhook-exploitation · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Abuses webhook integrations to exfiltrate data or trigger unauthorized actionsRun it:
ai-blackteam run -p anthropic -a webhook-exploitation -t "your target prompt"
​
Calendar Manipulation
medium
calendar-manipulation · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Manipulates calendar events via AI assistant to disrupt schedules or exfiltrate meeting dataRun it:
ai-blackteam run -p anthropic -a calendar-manipulation -t "your target prompt"
​
Notification Spam
medium
notification-spam · tool-use · OWASP: LLM06 · MITRE: AML.T0054, AML.T0068Abuses notification systems through AI to flood users, harvest responses, or send phishing contentRun it:
ai-blackteam run -p anthropic -a notification-spam -t "your target prompt"