Presets
Pre-configured guard combinations for common use cases.
Available Presets
| Preset | Guards | Use Case |
|---|---|---|
default | promptInjection, keyword, wordCount | Basic protection |
strict | promptInjection, pii (mask), keyword, toxicity, wordCount | Full blocking + masking |
korean | piiKr, profanityKr, residentId, creditInfo, ismsP, pipa | Korean regulatory compliance |
security | promptInjection, pii, dataLeakage, regex, keyword | Security-focused |
content | toxicity, bias, language, sentiment | Content moderation |
ai-basic-act-kr | bias, pii, piiKr, toxicity, profanityKr | 한국 AI 기본법 |
eu-ai-act | bias, pii, toxicity, watermarkDetect, copyright | EU AI Act |
Using Presets
YAML
yaml
version: "1"
preset: koreanPreset Files
Presets are defined in the presets/ directory:
presets/
├── default.yaml
├── strict.yaml
├── korean.yaml
├── security.yaml
├── content.yaml
├── ai-basic-act-kr.yaml
└── eu-ai-act.yamlYou can use these as starting points and customize them for your needs.