The Agent Skills Directory

[SAFE]: The skill provides official documentation and implementation patterns for Zoom's Translator API, serving as a developer guide for synchronous and asynchronous translation workflows.- [DATA_EXFILTRATION]: The skill documents the requirement for environment variables such as ZOOM_API_KEY and ZOOM_API_SECRET, as well as AWS credentials for S3-based batch jobs. It correctly advises that these secrets must remain server-side and provides clear warnings against browser-side exposure. Code examples for batch job submission use standard placeholders (e.g., AKIA...) for credentials.- [EXTERNAL_DOWNLOADS]: The skill references official Zoom documentation domains (developers.zoom.us) and the vendor's GitHub repository (github.com/zoom/ai-services-quickstart). As these resources are owned by the skill's author ('zoom'), they represent legitimate vendor infrastructure.- [COMMAND_EXECUTION]: No suspicious command execution patterns were identified. The provided Node.js examples use standard, well-known libraries like jsrsasign for JWT signing and the built-in node:crypto module for secure webhook signature verification.- [PROMPT_INJECTION]: The skill defines a standard translation interface where the agent ingests text or files for processing. While this represents a surface for indirect prompt injection (Category 8), the skill is a documentation layer for an external API and does not implement unsafe interpolation or autonomous decision-making based on the processed content.

translator