0570Reddit posts

Jev-kit: measured decision hooks and browser execution for Claude Code

A Reddit builder report links a public toolkit that uses Jev to judge selected tool calls, size subagents, and choose browser actions.

The original Reddit report and jonathanavis96/jev-kit source were inspected on October 1, 2026; airlock/client.py calls the native TypeSafe System One endpoint. Local rules prefilter tool calls, with Jev judging selected cases. The report separates preplanned Wikipedia clicks from tasks requiring navigation: Jev alone failed all six unguided tasks, while Sonnet planning plus Jev completed them. The kit also adds subagent-sizing and completion checks. These are small builder-run tests, and the guard fails open on provider errors.

Open source ↗ reddit.comcost-time-

jonathanavis96/jev-kitREADME ↗

jev-kit</h1>

<p>
  <strong>Everything you need to run TypeSafe's Jev with Claude Code:</strong><br/>
  guard your agent's tool calls, right-size its sub-agents, and wire Jev into
  search, browsing, review and more.
</p>

<p>
  <a href="LICENSE"><img alt="License: MIT" src="https://img.shields.io/badge/license-MIT-blue"></a>
  <img alt="tests" src="https://img.shields.io/badge/tests-1310%20passing-brightgreen">
  <img alt="Python 3.10+" src="https://img.shields.io/badge/python-3.10%2B-blue">
  <img alt="Platforms" src="https://img.shields.io/badge/platforms-Linux%20%7C%20WSL2%20%7C%20Windows-lightgrey">
  <img alt="Claude Code" src="https://img.shields.io/badge/Claude%20Code-PreToolUse%20hook-8A3FFC">
  <img alt="TypeSafe Jev" src="https://img.shields.io/badge/TypeSafe-Jev-0F9D58">
</p>

<p>
  <a href="#quickstart">Quickstart</a> &middot;
  <a href="#what-is-in-the-kit">What is in the kit</a> &middot;
  <a href="#jev-as-the-decision-maker-the-browser-agent">Browser result</a> &middot;
  <a href="#how-it-works">How it works</a> &middot;
  <a href="#measured-results">Measured results</a> &middot;
  <a href="#safety-model">Safety</a> &middot;
  <a href="#faq">FAQ</a>
</p>

<img src="docs/assets/hero.svg" alt="jev-kit: an agent's tool call passes a free code pre-filter that lets about 93% straight through; the rest are judged by Jev in about 0.3 s and then allowed, warned, rewritten or blocked. Any error lets the call through." width="100%">

</div>

> **Jev decides, Claude writes.** With Jev choosing each click, the browser
> agent's Claude bill was **0.0008 USD** a run against **0.1868 USD** with
> Sonnet choosing, at the same 9/9 success. [The numbers, and how they were
> taken.](#jev-as-the-decision-maker-the-browser-agent)

## What is Jev, and why a kit

Jev is a Type

Also filed under Agents & browsers

0226

Flight search with Browser Use★

Breaking: Browser Use + Jev = Ultrafast ⚡ Findings flights took 7s and cost only $0.0039 🤯 > new action space every step > DOM state space > small LLM fallback to type (this video is at 1x speed btw) Built a tiny open source browser agent. try it below ↓

@gregpr07 · Agents & browsers · ~$0.004 · ~7 s

0619

jevmem: automatic project memory for Claude Code

After every Claude Code message, Jev decides in ~0.3 s whether it holds a decision, rule or bug worth keeping, and jevmem saves it as one line in JEVMEM.md.

@jetwaniavinash · Agents & browsers · ~0.3 s

0618

Playwright + Jev: tests in plain language

Write Playwright steps like "add the most expensive item to the cart". Jev finds the right buttons and fields, fills forms and checks the result.

Andrey Popov · Agents & browsers

0617

toolgate: a tool-call firewall for agents

A Claude Code and MCP hook where Jev answers seven yes/no risk questions per tool call to allow, ask or deny. Usage and benchmark numbers are in the repo.

Jaz · Agents & browsers