Ask for input before execution if there are placeholders

[SaneBow]

I find that many times when I asked for a shell command, I forgot to specify every necessary concrete values, so the output command from GPT contains placeholders.
In those cases, I can't directly execute them, but need to copy the command, replace those placeholders, and paste to execute.

Why not check for placeholders and ask for concrete inputs before execution?
I've implemented it here.
If this is also in other people's favor, I will create a PR.

Here's how it looks like:

[SaneBow]

An alternative UX that support editing can be as the following:

Execute shell command? [y/i/e/N]

If user input e, we open the default shell editor (vi) to edit the command.
If user input i, we prompt for values for placeholders as proposed above.

[TheR1D]

Since your prompt does not specify a word or file name, GPT generates responses with placeholders. However, you can include all the necessary details in your prompt.

sgpt -s "count the occurance of 'python' in 'readme.md'"
grep -o 'python' readme.md | wc -l