Jev in production › Agent guardrails and approvals

jev-permission-gate

Jev replaces Claude Code's auto-mode tool-call classifier, judging whether each tool call is safe to run automatically. 2x faster decisions and zero unsafe allows across 93 labeled test cases (author).

Open on GitHub ↗

2x faster decisions and zero unsafe allows across 93 labeled test casesmeasured 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
2026-10-03
Found via
github
Repository
madisonrickert/jev-permission-gate
Stars
29
Forks
1
Last push
2026-10-04
Language
TypeScript
License
MIT

The README opens with

A Claude Code mod that puts TypeSafe's Jev in front of the auto mode classifier. In auto mode, for each tool call Claude Code would otherwise send to its built-in classifier, Jev answers eight yes/no questions in one request and the mod decides:

- allow when the call plausibly serves your request (at least 50%) and every risk check is at or below 10%. The built-in classifier doesn't run. - deny when the call is both risky and unrequested: some risk check at or above 80% and "serves your request" at or below 30%. Claude sees the reason. - defer everything else to the built-in classifier, unchanged.

Badge

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

Listed in Jev in production

Also used for agent guardrails and approvals