The Agent Skills Directory

[SAFE]: The skill's instructions and fixtures are designed for legitimate SDK integration tasks. All behaviors are consistent with its stated purpose of assisting developers with the Scandit MatrixScan Count features.\n- [EXTERNAL_DOWNLOADS]: The skill identifies and uses official Scandit NPM packages and documentation. These external resources are owned by the vendor ('scandit') and are necessary for the development tasks described.\n- [PROMPT_INJECTION]: Static analysis identified potential concealment patterns. These were reviewed and found to be benign developer settings (e.g., hiding UI elements in the scanner view) and helpfulness instructions (e.g., directing the agent to provide documentation links), rather than attempts to hide malicious behavior or override system prompts.\n- [CREDENTIALS_UNSAFE]: The skill uses safe placeholders for license keys and instructs users on where to obtain their own, adhering to standard secret management practices and preventing credential exposure.\n- [SAFE]: For processing external data such as packing slips, the skill correctly promotes the use of the SDK's structured BarcodeCountCaptureList API, which provides built-in validation and reduces the risk of logic errors during untrusted data handling.

matrixscan-count-capacitor