The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @zhoujinandrew/gpt-image-cli Node.js package from the npm registry to enable its core functionality.
[COMMAND_EXECUTION]: The skill utilizes the gpt-image-cli binary to perform complex image generation and editing tasks, as well as to manage user configuration and API endpoints.
[INDIRECT_PROMPT_INJECTION]: The skill implements a multi-round optimization workflow where the agent is instructed to use the Read tool to inspect generated PNG files and use the visual results to formulate subsequent prompts. This creates a surface where instructions hidden within generated images could theoretically influence the agent's logic.
1. Ingestion points: The agent reads generated PNG files from the ./refine/ directory.
1. Boundary markers: The skill provides a structured four-part template for prompt reconstruction to maintain consistency.
1. Capability inventory: The skill utilizes binary execution (gpt-image-cli) and file operations (Read, cp).
1. Sanitization: No specific sanitization or filtering logic is defined for the content of the generated images before they are processed by the agent.

gpt-image