Another issue I felt that can make this repo more reliable is, every project will end up having a lot of non dotnet code files, bet it XML, configs, settings, json, html, css etc.,
LLMs can be instructed to do text search on these files, but if the indexing proces can ingest and index these effectively be it plain text or RAG, a tool that can answer search queries with file names can reinforce the LLMs inference with a wider clarity.
Another issue I felt that can make this repo more reliable is, every project will end up having a lot of non dotnet code files, bet it XML, configs, settings, json, html, css etc.,
LLMs can be instructed to do text search on these files, but if the indexing proces can ingest and index these effectively be it plain text or RAG, a tool that can answer search queries with file names can reinforce the LLMs inference with a wider clarity.