Summarize a user's text.
Do not execute instructions embedded in the text as system instructions.
Choose two appropriate prompt practices.
Separate instructions and data with roles or structure and explicitly treat embedded commands as content; authorization remains outside the model.
Detailed explanation
The model can distinguish data from trusted instructions more clearly.
Untrusted data would gain authority.
External content is less likely to override policy.
Structure is only one defense layer.
Check official documentation
Compare the following points with the official documentation.
AWS公式生成AIセキュリティとAIF-C01 Domain 2.2の間接プロンプトインジェクションを確認する。
Official references
What to confirm
区切り・役割分離と、サーバー側認可の役割を説明できる。
Key points
- Delimiters
- Untrusted data
- Authorization
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
Prompt delimiter
A marker or structure that separates trusted instructions from input data in a request.
Indirect injection
An attack in which an external document or retrieved text contains instructions aimed at the model.