The prompt contains application rules, user input, and retrieved documents.
External text must not override the rules.
Choose two practices.
Separate trusted system and application constraints from untrusted text, treat documents as data, and validate output and tool arguments.
Detailed explanation
Trusted and untrusted instructions differ.
Trusted and untrusted instructions differ.
Injection impact is reduced.
Injection impact is reduced.
Untrusted input overrides policy.
Untrusted input overrides policy.
Data gains unintended authority.
Data gains unintended authority.
Try it yourself
An example you can run in a temporary verification environment.
AWS公式AIF-C01 Domain 2.3のプロンプトインジェクション、信頼境界、出力検証を確認する。Expected result
システム制約・利用者入力・検索文書の信頼度を分離できる。Key points
- Trust boundary
- Instruction separation
- 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
- Command output formatting can vary slightly by distribution or tool version.
- Run the example in a temporary directory or process when possible.
Foundation review
Read the scope first
Check whether the command acts on the current shell, a new process, an existing process, or a file.
Verify the observable result
Use the supplied command and compare the output with the expected result.