Feature Request: Enable Multimodal Capabilities by Supporting Image File Inputs

What version of Codex is running?

0.20.0

Which model were you using?

默认

What platform is your computer?

windows wsl

What steps can reproduce the bug?

Is your feature request related to a problem? Please describe.

Currently, the Codex CLI operates only on text-based input. It's
not possible to reference local image files in a prompt and have
the tool analyze their content. This limits its use cases, as
many development tasks involve visual context (e.g., UI design,
diagrams, error screenshots).

Describe the solution you'd like

I propose that the CLI be updated to recognize paths to local
image files (e.g., PNG, JPG). When such a path is detected in a
prompt, the CLI should read the image data and pass it to the
model along with the text prompt, enabling true multimodal
interaction.

Describe alternatives you've considered

The only alternative within the CLI is to manually describe the
image in text, which is inefficient and often loses critical
details.

Example Usage

A user should be able to run a command like this:

1 codex-cli "Based on the attached screenshot, please
generate the HTML and CSS for this component:
@/path/to/screenshots/login_form.png"
This would allow for a much richer and more efficient development
workflow. Thank you for considering this enhancement.

What is the expected behavior?

_No response_

What do you see instead?

_No response_

Additional information

_No response_