The Agent Skills Directory

[SAFE]: The skill documents the official @google/genai SDK for Google Gemini, providing legitimate patterns for text generation, multimodal processing, and function calling.- [SAFE]: It contains explicit security guidance in the 'CRITICAL' and 'RED FLAGS' sections, warning developers against hardcoding API keys and instructing them to use environment variables.- [SAFE]: Code examples demonstrate robust error handling, including checks for safety filter blocks (finishReason === "SAFETY") and rate limiting, which are essential for secure AI integration.- [SAFE]: The mention of BLOCK_NONE for safety settings in examples/advanced.md is correctly documented as a backend-specific feature of the Gemini API and is accompanied by a clear security caution regarding user-facing content.- [SAFE]: All identified dependencies (@google/genai, zod, zod-to-json-schema) are standard, well-known libraries from official registries.

ai-provider-google-gemini-sdk