What it does
A companion OpenCode plugin brings the checks into coding sessions. Checks send code and context to the selected provider; their bounded context is not a full program analysis, and cached results need deliberate refresh.
Agent tooling
Jevlint checks pieces of code against plain-language rules, with uncertain answers kept separate from passes and failures.
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
A companion OpenCode plugin brings the checks into coding sessions. Checks send code and context to the selected provider; their bounded context is not a full program analysis, and cached results need deliberate refresh.
Ask your developer to try one plain-language rule on a small part of your code. Jevlint sends pieces of that code to an outside AI service and reports likely violations. You need the checking program and a service access key. Keep ordinary tests too: these checks can miss problems beyond the code they see.