Refined the .check(callback) clause to #promptgun that allows you to check the LLM output, give it feedback and force it to iterate on it, with only a few lines of code. Now out in v1.1.2.