In recent years, the usage of large language models in code-related tasks has increased significantly, which has prompted the development of new tools to help developers and software development organizations optimize processes and increase their efficiency. Due to this rapid development, security threats and vulnerabilities have grown increasingly prevalent and researchers are prioritizing codeLMs security. In order to get a better understanding of such security problems, we present LLMs and codeLMs, explore strategies for attack that target codeLMs, review existing datasets, and highlight potential improvements for enhancing the security of these models.