What it is
A web tool and MCP server that reviews public GitHub pull requests, diffs or single files against Robert C. Martin's Clean Code. Built for developers who want a checkable, per-file verdict rather than prose alone.
How it uses Jev
For each code file, Jev answers a question set drawn from Clean Code, returning probabilities and scores rather than sentences. The question set is per file: test questions only on test paths, Boy Scout questions only on diffs. Luna then writes the review from those answers.
Primitives:choicescore
Technique worth stealing
Keep the question set as a single source of truth so prompt, payload and meters change together.
Try it
bun install, then bunx eve link and bun run dev; or use the live site and MCP endpoint.
Evidence
Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.
- Jev-centricyes0.80
- Shows a System One patternyes0.63
- Handles uncertaintyno0.04
- Measuredno0.06
- Runnableyes0.96
- Worth recommendingunclear0.40
- Model replicano0.05
- Problem scopescore on a 0–2 scale1.93
- About Jevyes0.97
Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 21 Sept 2026.