jev-style-adopt SKILL.md skillB

jev-style-adopt from lawrence3699/jev-style is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add lawrence3699/jev-style/jev-style-adopt@git:20260925.fc458b5

Pins git:20260925.fc458b5 in mdr.lock. Pin the fingerprint instead.

Audit grade
B, 16 of 17 checks passed
Kind
SKILL.md skill
Repository
lawrence3699/jev-style
Path
skills/jev-style-adopt/SKILL.md
Versions recorded
1
Latest version
git:20260925.fc458b5, committed Sep 25, 2026
Size
5,869 bytes
Repository stars
54
License
Apache-2.0

Find LLM API calls in a codebase that only return a decision (a category, yes/no, or a 1-5 rating) and move them to the local Jev-Style decision model safely - scan, pick candidates with the user, rewrite each as a typed noul/choice/score question with a confidence cascade back to the LLM, run in shadow mode, measure agreement with jev-style eval, then switch. Use when the user wants to cut LLM cost or latency for classification, routing, moderation, gating or scoring calls, or asks which LLM…

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Replace decision-only LLM calls with Jev-Style

Many codebases send a hosted LLM a prompt and then parse one word, a boolean or a number out of the
reply. Those calls pay for text generation to get a label. Jev-Style answers the same question
locally, in one forward pass, with a probability per option. This skill moves them over without
changing behaviour blindly: **scan -> choose -> rewrite with a cascade -> shadow -> measure -> switch**.

Never delete the LLM path in the first change. Every step below is reversible.

## 1. Scan

```sh
python3 <this-skill-dir>/scripts/find_decisions.py path/to/src --json /tmp/decision_calls.json
```

It lists LLM call sites whose prompt or parsing looks decision-shaped (classify / "yes or no" /
"respond with only" / rate 1-5 / enum or bool schemas / tiny `max_tokens` / `.strip().lower() == "yes"`)
with a guessed type. It is a heuristic: read every hit. Drop calls that need generated text, world
knowledge not in the input, multi-step math, or tool use.

## 2. Choose with the user

Show a short table: file:line, what it decides, guessed type, how often it runs (ask, or look for
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add lawrence3699/jev-style/jev-style-adopt@sha256:0fdd70f2e992edd9

The line at the top of this page pins the label git:20260925.fc458b5, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_ynmmqgg7gm4qw6i3.svg)](https://markdownregistry.com/a/art_ynmmqgg7gm4qw6i3)

1 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 1 version of this file, committed on Sep 25, 2026.

versioncommittedcommitsizeaudit
git:20260925.fc458b5 latest2026-09-25 fc458b5 5,869 BB view

Audit of the latest version

B  16 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • fail: No link to a raw IP address (matched: http://127.0.0.1:8765)
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (5869 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No script tag

Starting a new one? Use a template for SKILL.md on mdpantry, which passes every one of these checks and says what to change in it.

Source

GitHub

lawrence3699/jev-style · 54 stars · license Apache-2.0 · pushed 2026-11-04 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_ynmmqgg7gm4qw6i3
GET https://markdownregistry.com/api/v1/resolve?ref=lawrence3699/jev-style/jev-style-adopt
GET https://markdownregistry.com/api/v1/blob/0fdd70f2e992edd9d2f5b5b481647c5916f9f7f3c7532f51cf97329a76914ee4

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from lawrence3699/jev-style

jev-style-eval SKILL.md skill
lawrence3699/jev-style · skills/jev-style-eval/SKILL.md · Measure the Jev-Style local decision model on the user's own labelled examples before relying on it - accuracy, Brier…
git:20260925.fc458b5 · audit B · 54 stars
jev-style-guard SKILL.md skill
lawrence3699/jev-style · skills/jev-style-guard/SKILL.md · Install the Jev-Style guard, a Claude Code PreToolUse hook that has a local decision model check every Bash, Write…
git:20261104.4c49bf7 · audit B · 54 stars
jev-style-mcp SKILL.md skill
lawrence3699/jev-style · skills/jev-style-mcp/SKILL.md · Connect the Jev-Style local decision model to an AI coding agent as an MCP server (tools decide, noul, choice, score…
git:20261104.4c49bf7 · audit B · 54 stars
jev-style-serve SKILL.md skill
lawrence3699/jev-style · skills/jev-style-serve/SKILL.md · Install and start the Jev-Style local decision server (systemone-compatible API + Playground at http://127.0.0.1:8765)…
git:20261104.4c49bf7 · audit B · 54 stars
jev-style SKILL.md skill
lawrence3699/jev-style · skills/jev-style/SKILL.md · Call the local Jev-Style decision model (systemone-compatible API, default http://127.0.0.1:8765) to answer typed…
git:20261104.4c49bf7 · audit B · 54 stars

Every file in lawrence3699/jev-style

Browse by kind, by grade B, or by owner.