Structured debugging for backend-as-a-service applications

[krabat-l]

When AI agents debug Appwrite-backed apps, issues can be in the SDK call, the function, the permissions, or the database. I built a methodology that enforces systematic investigation: claude-debug

Reproduce first, isolate by logging at each layer, confirm root cause, then fix.

https://github.com/krabat-l/claude-debug/blob/master/rules/debug.md