jevbooks

← All projects

jev-semgrep

uehaj/jev-semgrep

Grep by meaning across languages using Jev probabilities.

Scoring & Ranking81%Runner-up: Classification & Routing

What it is

semgrep is a command-line tool that finds lines by what they mean, not by regular expressions. It is useful for mixed-language logs, ticket dumps, and teams querying in different languages.

How it uses Jev

For each line, semgrep asks Jev a typed question: does this line match the given meaning? Jev returns a probability, and semgrep applies a threshold. Meanings combine with AND, OR, and NOT using boolean operations on these probabilities.

Primitives:score

Technique worth stealing

Use calibrated probabilities from a cross-encoder to enable logical AND/OR/NOT on meanings without index building.

Try it

Run npx @uehaj/semgrep -n -e "customer is angry or frustrated" tickets.txt after setting TYPESAFE_API_KEY.

View on GitHub

judged by Jevjev-1.13.0

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-centricyes0.95
  • Shows a System One patternyes0.75
  • Handles uncertaintyyes0.64
  • Measuredno0.14
  • Runnableyes0.92
  • Worth recommendingunclear0.59
  • Model replicano0.04
  • Problem scopescore on a 0–2 scale1.36
  • About Jevyes0.98

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.