Text checked against seven AI safety guardrails
The scenario runs texts from Google Sheets through seven checks — bypass attempts, personal data, secrets, unsafe content, links, topic and stop-words — scores the risk and scrubs what's dangerous.
- Seven safety checks in one scenario
- Personal data and keys are scrubbed automatically
- AI scores the risk of each text
- Test an assistant before it goes live to clients
Guardrails
Made with Scriptera