codex-cli run_in_container.sh appears stale relative to devcontainer firewall setup

What happened?

codex-cli/scripts/run_in_container.sh appears to be stale relative to the current devcontainer setup.

A few examples:

• The script runs an image named codex, but the current devcontainer docs build images from .devcontainer with a configurable image name.
• The script calls /usr/local/bin/init_firewall.sh and then removes that same path.
• The current secure devcontainer copies the firewall helper to /usr/local/bin/init-firewall.sh with a hyphen, and .devcontainer/post-start.sh invokes that hyphenated path.
• The usage examples describe passing a shell COMMAND, but the script ultimately appends the arguments to codex --sandbox workspace-write --ask-for-approval on-request, which treats them as Codex prompt input rather than executing a shell command directly.

Expected behavior

Either the helper should be updated to match the current devcontainer image/firewall layout and command semantics, or the stale helper should be removed/documented as obsolete if it is no longer part of the supported container workflow.

Additional context

I searched open issues and PRs for terms including run_in_container.sh, init_firewall.sh, and the container helper path. I found an unrelated secure-devcontainer DNS issue mentioning firewall setup, but did not find an existing open item for this stale codex-cli/scripts/run_in_container.sh helper.