ai-security - Assess AI and LLM security risks
Scans AI/ML systems for prompt injection, jailbreaks, model inversion, data poisoning, and agent tool abuse, with MITRE ATLAS mapping.
Tags
Updated: 2026-10-03Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Scan prompt injection signatures
- Assess jailbreak resistance
- Score model inversion risk
- Score data poisoning risk
- Detect agent tool abuse
- Map findings to MITRE ATLAS
- Recommend guardrail controls
Inputs
- Test prompts
- Prompt test file
- Target type
- Access level
- Threat scope
- Testing authorization
Outputs
- Injection findings
- Risk scores
- ATLAS technique mappings
- Guardrail recommendations
- Scanner exit status
Requirements
- Python 3
- AI threat scanner script
- Written authorization for gray-box access
- Written authorization for white-box access