The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of local Python scripts, specifically scripts/image_gen.py and remove_chroma_key.py, to handle interaction with external APIs and perform image post-processing (such as background removal).
[EXTERNAL_DOWNLOADS]: The skill documentation instructs users to install standard, well-known Python packages (openai, pillow) via the uv package manager and involves network communication with official OpenAI API endpoints.
[CREDENTIALS_UNSAFE]: The skill requires an OPENAI_API_KEY for its CLI fallback mode but correctly instructs the user to provide this via environment variables, avoiding the risk of hardcoded credentials.
[PROMPT_INJECTION]: The skill manages potential indirect prompt injection risks through a structured ingestion and normalization workflow.
Ingestion points: User-provided text prompts and reference images are ingested for processing (documented in SKILL.md).
Boundary markers: The skill implements a 'Shared prompt schema' that uses explicit labels to structure the user input and separate it from internal instructions.
Capability inventory: The skill has the capability to write files to the workspace and make network calls via the provided CLI scripts.
Sanitization: A prompt augmentation and normalization process is used to validate and structure user input before it is passed to the generation model, reducing the surface area for adversarial input.

imagegen