domains / families / security / ai-ml-security

ai-ml-security

separate

AI/ML security playbook. Use when assessing model supply chain attacks (pickle RCE, poisoned weights), adversarial examples, model poisoning, model stealing, data privacy attacks (membership inference, model inversion), and autonomous agent security risks.

When to use

• model supply chain attacks
• pickle RCE
• poisoned weights
• adversarial examples
• model poisoning
• model stealing
• data privacy attacks
• autonomous agent security risks

When not to use

• LLM-specific prompt injection
• jailbreaking
• tool abuse techniques
• general deserialization attack patterns
• pip/npm package confusion

Skill page

Source