A team wants to apply the same supported Bedrock content safeguards to text from a non-Bedrock model. Which API capability should it evaluate?

1. Use ApplyGuardrail independently for the supported input and output assessment workflow.
2. Use IAM policy simulation as a substitute for text safety assessment.
3. Copy the external model API key into the guardrail prompt.
4. Assume guardrails can only be used with one hardcoded foundation model.