A question about custom language support

[ghost]

If a language, for example, is defined in both eepp source code and Adept.json, then which takes precedent?

[SpartanJ]

It always prioritizes any loaded language, so it basically will try to use Adept.json. The idea is to let the user do their own modifications, also you can export the language definition to the json format to extend it (I guess you know since you mentioned the json file) with the export-lang CLI option:
ecode --export-lang=cpp --export-lang-path=./cpp.json

[ghost]

How to export all of the languages?

[SpartanJ]

You can achieve it with some script by getting the language names from the ecode --health output, this should work on Linux and macOS:

#!/bin/bash
ecode --health --health-format=ascii | tail -n +2 | while read -r line; do
	lang_col="${line%%✓*}"
	lang=$(echo "$lang_col" | sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//')  # Trim whitespace
	ecode --export-lang="$lang" --export-lang-path=./langs/"$lang".json
done