[vielite1310]

I would like to be enlightened myself if RPI,BMAD or any spec-driven approaches actually worked for any mid/large scale projects, without wasting millions of tokens of course :)

[rsrsrs86]

Real spec driven uses no tokens