Post #15974 203 Aug 31, 2026, 16:06 UTC I built a local ~150M prompt quality scorer you can put in front of your LLM/agent (open source)https://redd.it/1w3g1yc@r_CoolGithubProjects