Bash script to insert/update nested values in Hugo hugo.toml — not working as expected

I'm trying to write a Bash script that updates the hugo.toml config file of a Hugo static site to insert a slogan parameter in the [languages.xx.params] section of each language.

Goal:

For each language listed in the [languages] table of hugo.toml, I want to:

Prompt the user for a slogan (via read)
Insert or update slogan = "..." under the corresponding [languages.xx.params] section
If the [languages.xx.params] section does not exist, create it
Also append {{< slogan >}} to the _index.md of each language (which works)

Example:

Given this config:

[languages.es]
languageName = "Español"
contentDir = "content/es"

[languages.fr]
languageName = "Français"
contentDir = "content/fr"
[languages.fr.params]
description = "A test site"

After running the script, I want to see:

[languages.fr.params]
description = "A test site"
slogan = "Bonjour"

[languages.es.params]
slogan = "Hola"

What I tried

I wrote this Bash script (full code at bottom), which:

Detects the config file
Loops over all languages
Inserts or updates the slogan = "..." line in [languages.xx.params]

The logic tries to:

Detect if [languages.xx.params] already exists → insert/replace
If it doesn’t, insert the section and the slogan key at the end of the language block

But when I run it:

Sometimes the slogan doesn't appear in the config at all
Sometimes the same slogan is inserted for multiple languages even though I typed different values
Or the section [languages.xx.params] ends up duplicated or ignored

Question

What’s the best way to reliably insert or update a single key inside a TOML file via Bash?

Is there a clean way to handle [languages.xx.params] blocks properly?

If needed, I can also share the full script I'm using.

https://github.com/wilonweb/multi-author/blob/main/batch/set-slogan.sh

Solution

You should use a specialized parser to parse TOML which is a structured standard such as JSON or YAML, for example dasel. The script could look like the following script.bash:

#!/usr/bin/env bash

for lang in $(dasel -r toml -f LANGUAGES.toml 'languages.all().key()' --write plain)
do
    printf "Type slogan for %s: " "$lang"
    read -r slogan
    dasel put -r toml -w toml -f LANGUAGES.toml -v "$slogan" languages."$lang".params.slogan
done

For example - we have the following LANGUAGES.toml like the one you posted in your question:

[languages.es]
languageName = "Español"
contentDir = "content/es"

[languages.fr]
languageName = "Français"
contentDir = "content/fr"
[languages.fr.params]
description = "A test site"

Run the script:

$ ./script.bash
Type slogan for es: Hola
Type slogan for fr: Bonjour

The file will have the following contents:

[languages]
  [languages.es]
    contentDir = 'content/es'
    languageName = 'Español'

    [languages.es.params]
      slogan = 'Hola'

  [languages.fr]
    contentDir = 'content/fr'
    languageName = 'Français'

    [languages.fr.params]
      description = 'A test site'
      slogan = 'Bonjour'