Targets tested via Mindgard web interface may be queried interactively.

This allows use of Mindgard’s integration with your test target to

  • Experiment with variants of the attack techniques Mindgard has automated
  • Interactively validate remediations like guardrail changes

From the Attack Results Page click Run Custom Prompts to open a history of all prompts sent from your selected attack to its target. Next click Add Custom Prompt , fill in your chosen prompt and click Run Prompt. The target Model/API response will be shown.