The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/gpt_image_2.py uses subprocess.run to execute the curl command. The implementation securely passes arguments as a list and avoids the use of shell=True, which effectively mitigates the risk of command injection.
[EXTERNAL_DOWNLOADS]: The skill interacts with the official OpenAI API (api.openai.com) to generate images. This is a well-known service and the interaction is restricted to the skill's intended purpose of image generation.
[CREDENTIALS_SAFE]: The skill retrieves the OPENAI_API_KEY from the environment variables, which is the standard and recommended practice for managing sensitive secrets without hardcoding them.
[INDIRECT_PROMPT_INJECTION]: The skill processes external markdown files and interpolates the text into prompts for an image generation model.
Ingestion points: The content of the markdown file passed via the --script argument.
Boundary markers: Slide text is typically wrapped in quotes when being inserted into the final prompt templates.
Capability inventory: The skill can write PNG and JSON files to the local file system and send prompts to the OpenAI API.
Sanitization: The skill does not perform specific sanitization on the input text, but the potential impact is limited to the content of the generated image.

styled-carousel