slopcheck-jev

A prose linter that catches AI writing tells. Regex settles the 18 a pattern can settle. Jev takes the 15 that need reading, as 15 Nouls in one call, 604 ms median. It ships as a Claude Code `Stop` hook that scores Claude's own output after every turn and warns rather than blocks.

GitHub

License: MITLanguage: TypeScriptLast push: 9/19/2026#cli-git-gates#typed-decisions

What Jev Does in this Architecture

**Deleting a feature made it more accurate.** The first build made a second Jev call to pin each tell to a sentence. Removing it took precision from 0.80 to 0.95 and latency from 1,365 ms to 575 ms. Jev returns typed answers rather than text, so a quote has to come from a second round of typed questions about lines, and that round was where every defect lived. The README documents the two intermediate designs that were worse.

Key Architecture Benefit

Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.

Reviewed Source Code Evidence

Verified fixed-commit source files demonstrating genuine Jev integration:

harshpuri84/slopcheck-jev/blob/f37c440b84aaa09e0c2de5f441f19a32f8711781/README.md

harshpuri84/slopcheck-jev/blob/f37c440b84aaa09e0c2de5f441f19a32f8711781/slopcheck/gate.py

harshpuri84/slopcheck-jev/blob/f37c440b84aaa09e0c2de5f441f19a32f8711781/slopcheck/judge.py

Review & Benchmark Note: Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.

Clone Repository

git clone https://github.com/harshpuri84/slopcheck-jev.git