The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it accepts and processes user-provided text prompts and image files that are subsequently transmitted to external AI models for processing. Malicious instructions embedded within these inputs could influence the agent's behavior. The risk is heightened by the skill's capability to perform network operations and write files to the local disk. Ingestion points: The prompt argument and the optional input_image file path in scripts/generate_image.py. Boundary markers: None identified; inputs are passed to the API without delimiters or instruction-ignore headers. Capability inventory: The skill can perform network POST requests to openrouter.ai and binary file write operations on the local filesystem via scripts/generate_image.py. Sanitization: None; the script does not validate or filter the content of the input prompt or image before transmission.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the requests library and makes network requests to openrouter.ai to access image generation and editing services.
[COMMAND_EXECUTION]: The documentation instructs the user to execute a Python script (scripts/generate_image.py) that performs network communication and file system modifications.

generate-image