The Agent Skills Directory

[SAFE]: The skill provides standard documentation and Python code snippets for using the official OpenAI SDK to process images via vision-capable models.
[SAFE]: External references to OpenAI and other providers (DeepSeek, Zhipu, etc.) target well-known and established services.
[PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface as it facilitates processing data from external images.
Ingestion points: Untrusted image data enters the context via image_url blocks in the messages array.
Boundary markers: The examples do not include explicit instructions or delimiters to mitigate the impact of text found within images.
Capability inventory: The skill enables full interaction with LLM providers for visual analysis and extraction.
Sanitization: No sanitization or safety checks are applied to the image data or its source before being sent to the model provider.

openai-vision