How to parse a string of a string in Ruby?

Use regex

s = '"token"'
str = s[/\A"(.+)"\Z/, 1]
puts str # => token

Try:

JSON.parse(response.body) rescue response.body[1..-2]

Why not sniff for the normal JSON open/close characters [...] and {...} used to define objects?

Here's how objects look converted to JSON:

require 'json'

puts JSON[{'foo' => 1}]
puts JSON[[1, 2]]
# >> {"foo":1}
# >> [1,2]

Checking to make sure those are valid string representations:

JSON['{"foo":1}'] # => {"foo"=>1}
JSON['[1,2]'] # => [1, 2]

Here's how I'd write some code to sniff whether the string passed in looks like a valid JSON string, or just a string with wrapping double-quotes:

require 'json'

def get_object(str)
  if str[/^[{\[].+[}\]]$/]
    JSON[str]
  else
    str[1..-2]
  end
end

get_object('{"foo":1}') # => {"foo"=>1}
get_object('{"foo":1}').class # => Hash
get_object('[1,2]') # => [1, 2]
get_object('[1,2]').class # => Array
get_object("'bar'") # => "bar"

Notice that on the last line the string 'bar' that's enclosed in single-quotes, is returned as a bare string, i.e., there is no wrapping set of single-quotes so the string is no longer quoted.

It's possible you'd receive a string with white-space leading or following the payload. If so, do something like:

def get_object(str)
  if str.strip[/^[{\[].+[}\]]$/]
    JSON[str]
  else
    str.strip[1..-2]
  end
end

get_object('  {"foo":1}') # => {"foo"=>1}
get_object('  [1,2]  ') # => [1, 2]
get_object(" 'bar'") # => "bar"