to design a questionnaire, I would recommend to first consider the categories you would like to assess by using the questionnaire. Subsequently, you could design questions that, according to your opinion, best represent the individual categories. You could then present these questions to a representative sample of participants and ask them to answer the questions. Once you have collected all the answers, you could apply a factor analysis (principal components analysis) that assists you in choosing individual questions, i.e. for each category you like to assess, you try to find the questions that best represent this category. Finally, you would be able to apply the questionnaire with the chosen questions. To get an estimate on external validity, you could then compare the results of the questionnaire with other indicators, e.g. how well people communicate in business English according to a rating of native speakers in the business sector.
I agree with Rainer's observations. Develop your questionnaire based first on what you want to research, then what form to do want the questionnaire to have, and finally, what do you want to do with the results. Reliability to me means that the questionnaire can be administered in different settings, with similar results. Take part of the questionnaire, have others administer it, and then compare the results. Validity is a little more complex. You will be the best judge of validity. Did it measure what you intended it to measure? Was it similar to other questionnaires that have been used in other studies? In the end, did the questionnaire measure what you wanted it to measure? All of this can be done in a pilot study format. Once you have that information, then revise the questionnaire, as needed.