The Agent Skills Directory

Secret Management Best Practices: The skill provides explicit guidance on how to securely handle sensitive credentials like the OPENAI_API_KEY. It instructs users to utilize environment variables and warns against pasting keys directly into the chat interface, which aligns with industry security standards for credential handling.
Standard External Dependencies: The skill identifies and uses well-known, trusted Python packages such as openai and pillow. These are standard tools for interacting with APIs and processing images, and the skill provides instructions for installing them via established package managers like uv.
Structured Prompt Ingestion: The skill implements a structured approach to processing user-supplied prompts. In scripts/image_gen.py, user input is interpolated into a template with clear labels such as 'Primary request:' and 'Constraints:'. This practice helps the underlying AI model distinguish between instructional context and data, which is a key consideration for mitigating indirect prompt injection.
Defined Execution Scope: The skill manages its activities within clear boundaries. It restricts file writing to specific project directories like output/imagegen/ and only initiates network connections to the OpenAI API, which is a well-known service required for the skill's primary function.

imagegen