Meta researchers introduced TestGen-LLM, an automated tool aimed at improving test coverage in software engineering. While not releasing the code, they inspired others to implement it, resulting in the open-source Cover-Agent. The process involves generating tests with Generative AI, filtering for functionality and coverage, and manual reviews. Challenges included the nuances of different programming languages and the need for user input for better results. Cover-Agent aims to automate the testing process but still requires developer oversight. Future goals include enhancing the tool’s capabilities and encouraging collaboration to advance generative AI for testing. Ultimately, the vision for Cover-Agent is to automate test suite enhancements for codebases.
https://www.codium.ai/blog/we-created-the-first-open-source-implementation-of-metas-testgen-llm/