The code currently fetches all CVEs from NVD api. However, when it is run a second time, it could be using the modified time to update the pandas data frame instead of fetching all and overwriting. This would make the code much more efficient and as the CVEs grow, this will begin to mater a lot.