A generative assistant calls an inventory API when needed.
Control which external operations the model can use.
Which statement about agentic AI tool use is appropriate?
An agent can select a defined tool, call it with a validated schema, and use the result in a later response; permissions and validation remain necessary.
Detailed explanation
A tool definition and execution permission allow an agent to call an external function.
Tool results can be stale or erroneous and should be validated.
Tool invocation is different from model training.
Tool access must be constrained by least privilege.
Check official documentation
Compare the following points with the official documentation.
Amazon Bedrock AgentsまたはBedrock AgentCoreの公式ドキュメントで、ツール呼び出しと権限の説明を確認する。
Official references
What to confirm
ツール定義・権限・入力出力の検証を、モデルの再学習と区別できる。
Key points
- Tool invocation
- Least privilege
- Input and output validation
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Action authority
An agent may propose an action, but a separate policy boundary should decide whether it can run.
Failure containment
Timeouts, idempotency, and bounded retries limit damage when a tool or model behaves unexpectedly.