# Planned Improvements - add more output formats - Text, Markdown, maybe others. - add a test suite to ensure the tool works as expected in a variety of scenarios. - allow the user to add extra file exclusions, or allow files that are excluded by default to be included. - add the ability to check for updates and update the tool (or at least notify the user that an update is available and where to get). - actually remove comments from the generated XML file. Perhaps add a flag to allow the user to choose whether to include comments or not. Looking at using 'tree-sitter' to parse the code and remove comments. Need to think about needed comments - like for linters and such, we may not want them removed? perhaps start by removing block comments and all docstrings? Then again, it is always good to have comments in the returned code from the LLM. Needs more thought. **If we add this, remember to re-add the note on comment removal to the output file `` node.** - ensure that the tool works on Windows, Linux, and macOS. It does work great on all 3 at this current code state, but we need to develop a test suite and get the CI pipeline working to ensure that it continues to work on all 3. - allow to work with non-git repositories (local only obviously). - see if we can find a library to tokenize other models such as `Claude` or `Gemini` - change file encoding to UTF-8 for included files, this is to ensure that the XML file is valid and can be read by other tools and specifically LLM's who generally prefer UTF-8. - allow user to add custom metadata to the XML file, this could be used to store information about the repository, such as the name, description, extra instructions, etc. Would again be once the TOML file is implemented. - ignore `dotfiles` by default, but allow the user to include them if they want. - Add secret-checking to the tool, to ensure that no secrets are included in the output XML file. Hopefully this can be done with a library, but may need to write our own checks.