Analysis of the Precision of Large Language Models in the Identification of Security Vulnerabilities and Weaknesses in Generated Code