matrixscan-batch-web
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill includes instructions that direct the agent to prioritize provided reference materials over its internal training data to ensure technical accuracy regarding API versions. It also specifies behavioral constraints, such as requiring the agent to fetch documentation before responding and providing source links to the user. These are standard prompt engineering techniques for technical support skills designed to prevent hallucinations.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for installing official Scandit software packages from the NPM registry and fetching engine assets from the Scandit-maintained directory on the JSDelivr CDN. These are legitimate resources required for the functionality of the SDK.
- [COMMAND_EXECUTION]: The documentation includes standard development commands (e.g.,
npm install) and configuration examples for server headers (COOP/COEP) and Vite middleware, which are necessary for the multithreading requirements of the SDK.
Audit Metadata