Why was local score calculation removed? + small transparency ask

[sebastiantuyu]

this is a very nice project, it's really interesting for better best-practices addition!

couple questions regarding transparency & local usage:

i noticed calculateScoreLocally was introduced in this commit which was a super nice thing, but i also noticed it was removed in this other commit. i also noticed the llms.txt used to mention the score was locally computed but it doesn't anymore, which is what made me curious. i guess my questions/suggestions are:

• why no local score calculation?
• i believe that to be transparent this should be explicitly mentioned somewhere, at least a simple section noting the score is only calculated through the API which i belive is a deciding point.

[aidenybai]

thanks for raising this! telemetry is actually explicitly mentioned here: https://github.com/millionco/react-doctor#telemetry

on the local score calculation: we removed it because scoring is something we tweak pretty frequently right now. React Doctor is still relatively new, and we care a lot about the score being accurate / realistic rather than stale or misleading

the rules themselves are part of the package, but the scoring methodology can change as we fix false positives, reweight issues, or learn which diagnostics are actually most important in real-world codebases. because of that, we currently prefer no score over a bad score

that said, if you’re at a company with specific security requirements here, we do have an enterprise option and are happy to work through what that setup could look like.

agree we can make this clearer in docs too. the telemetry section exists, but we should probably explicitly mention that the score is calculated through the API today. appreciate the thoughtful issue!