domains / families / security / phy-content-safety-guard

phy-content-safety-guard

separate

Dual-layer AI content guardrail with red-team test methodology

When to use

• Intercepting outbound messages
• Evaluating AI agent outputs
• Red-team test methodology
• Guarding the output layer
• Preventing prompt injection leaks
• Blocking sensitive content

When not to use

• Input filtering

Skill page

Source