The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill communicates with the OpenRouter API service to perform image generation and editing tasks.
The script scripts/generate_image.py makes POST requests to https://openrouter.ai/api/v1/chat/completions.
[DATA_EXFILTRATION]: The skill accesses local configuration files to retrieve API credentials.
The check_env_file function in scripts/generate_image.py searches the current and parent directories for a .env file to read the OPENROUTER_API_KEY.
While this involves reading a sensitive file path and sending the content to an external API, it is the intended primary purpose of the skill for authentication and follows standard secret management practices.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied prompts and image data which are subsequently passed to an external AI model.
Ingestion points: User-provided prompt argument and the contents of the input_image file in scripts/generate_image.py.
Boundary markers: The data is transmitted within a structured JSON object but lacks explicit delimiters or instructions to the model to ignore embedded commands.
Capability inventory: The skill has network access (via the requests library) and file system access (reading inputs and writing the generated generated_image.png).
Sanitization: No specific sanitization or filtering is applied to the user input before transmission to the API.

generate-image