Gauntlet test suite
Attack the implementation with the smallest risk-complete hostile test set, then continue to the real outcome.
Attack the implementation with the smallest risk-complete hostile test set, then continue to the real outcome.
Nothing: your AI starts it with your words.
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this workflow by what it does and runs it.
Say to your AI: “Use Trillion to attack the implementation with the smallest risk-complete hostile test set, then continue to the real outcome”
What your AI calls: start_workflow {"workflow_id":"workflow:skill-gauntlet-test-suite"}