id-capture-android

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides integration guides and API references for the Scandit ID Capture SDK on Android. All external links point to official documentation, the vendor's licensing portal, or the standard Maven Central registry (Sonatype). The instructions regarding background thread execution and disabling the scanner during result processing are standard mobile development practices to ensure UI stability and prevent redundant scanning of the same document.
  • [INDIRECT_PROMPT_INJECTION]: The skill implements a data ingestion surface where identity documents (passports, driver's licenses, and barcodes) are scanned and converted into structured data. As with any system processing external data, the content of these documents is untrusted and could potentially contain adversarial content.
  • Ingestion points: The CapturedId result object received in the onIdCaptured callback across SKILL.md and references/integration.md.
  • Boundary markers: None identified; the skill focuses on technical API integration.
  • Capability inventory: The skill provides camera-based data extraction and result mapping; it does not include tools for arbitrary code execution or unauthorized network communication.
  • Sanitization: The code samples demonstrate how to display and log captured fields; developers using this skill are responsible for sanitizing document data before including it in broader application logic or downstream AI prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 07:18 AM
Security Audit — agent-trust-hub — id-capture-android