Jev in production › Agent guardrails and approvals

jev-kit

Claude Code PreToolUse hook where Jev judges tool calls its rules table leaves open, then allows, warns, rewrites or blocks; Jev also picks each browser-agent click, at 0.0008 USD of Claude spend a run vs 0.1868 USD with Sonnet (author).

Open on GitHub ↗

0.0008 USDmeasured against a baseline, as published by the source
Use
Agent guardrails and approvals
Industry
Developer tools
Form
Open-source tool
Stage
In production
Plugs into
Claude Code
Listed
2026-09-24
Found via
github
Repository
jonathanavis96/jev-kit
Stars
63
Forks
7
Last push
2026-10-02
Language
Python
License
MIT

The README opens with

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

Quickstart · What is in the kit · Browser result · How it works · Measured results · Safety · FAQ

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.

Badge

For the project's own README, linking back here:

Listed in Jev in production

Also used for agent guardrails and approvals