The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Python subprocess module to orchestrate its workflow. It executes the codex CLI for image generation and login status checks, and calls its own internal helper scripts (e.g., update_index.py, check_prompts.py) to maintain the style library and validate inputs. These operations are essential for the skill's stated purpose and are implemented using safe practices.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text (articles) to generate descriptive scenes for image creation. While this input is eventually used to form prompts for an image generator, the skill implements a structured pipeline with defined templates and validation steps (e.g., check_prompts.py) that serve as boundaries to prevent unintended agent behavior.
[DATA_EXPOSURE]: The skill manages user configuration and custom styles in a local directory (~/.image-render/). It also accesses session logs from the Codex CLI (~/.codex/sessions) to perform integrity checks, ensuring generated images match the intended prompts. These file operations are scoped to the application's own data and do not target sensitive system or user files.

image-render