Investigate if calling LLM from xscan is necessary
In the above MR, we plan to test if calling LLM from xscan is necessary.
The steps: The steps should look like this: 1.feed the code generation LLM with a bare list of libraries 2.feed the code generation LLM with a list of libraries with LLM generated description 3.compare the responses
the xscan report is used here: it is inserted by this function https://gitlab.com/gitlab-org/gitlab/-/blob/17ca19aa67d97f254963758e77b682e9302ff4[…]0c/ee/lib/code_suggestions/prompts/code_generation/anthropic.rb
Edited by Tian Gao