Dual Scan
AI-generatedSummary
Dual Scan prompt and response content
Inputs
- scanMode — Select synchronous or asynchronous scan mode
- promptContent (required) — The prompt content to scan
- responseContent (required) — The response content to scan
- context — Optional context for grounding validation (up to 100K characters)
- additionalOptions — Additional configuration options for the scan
Output shape
array of scan result objects
Each item contains operation metadata, transactionId, and detection details such as maskedContent or dlpDetected; successful execution returns a non‑empty array, errors may be swallowed or thrown depending on continueOnFail
Examples
Example 1: Detect data leakage in a user prompt and its AI response
Set Scan Mode to async, provide Prompt Content and Response Content, optionally add Context, and set AI Profile Override to "production-profile"
Example 2: Batch scan multiple items
Add up to five items under Batch Items, choose itemType=Prompt or Response, set Scan Mode=sync, and leave Additional Options at defaults
Example 3: Mask sensitive data in a large response
Enter a large Response Content, enable Async mode, set Max Polling Duration to 600000 ms, and optionally increase Max Retries