The DirectorySearchTool docs do not sufficiently articulate that this tool is ONLY indexing the file name and associated metadata and not RAG’ing the contents of those files.
If this is the intended purpose of this tool then it is not working atm and as such it a bug. This ic very evident if the SQL db is interogated.
In order to RAG over files then specific tools like PDFSearchTool is needed.
Could the associated doc page be updated, please, to ensure this confusion is avoided?