The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses argparse to handle user input for image prompts and file paths. The execution is limited to standard Python file operations and HTTP requests via the requests library. No arbitrary shell execution or dangerous system commands were detected.\n- [DATA_EXPOSURE]: The script reads from a .env file to retrieve the OPENROUTER_API_KEY. This is a documented best practice for secret management in development environments. The analysis confirmed that this access is scoped to the specific key and does not attempt to exfiltrate other sensitive files like SSH keys or AWS credentials.\n- [EXTERNAL_DOWNLOADS]: The skill communicates with openrouter.ai, which is a well-known service for accessing various AI models. The API calls are made over HTTPS, and the response data (base64-encoded images) is processed locally to save files. This behavior is consistent with the skill's primary purpose.\n- [PROMPT_INJECTION]: While the skill takes a user-provided prompt, it is passed as a data payload to an external image generation model. The skill does not possess sensitive capabilities (like database access or system administration) that would be vulnerable to standard prompt injection attacks. Standard LLM safety filters from the providers (Google, Black Forest Labs) would apply at the API level.

generate-image