Jev in production › Agent guardrails and approvals

jev-blindspot

Flags blind spots in a submitted prompt before a coding agent acts, using a TypeSafe key inside a side panel for Claude Code and Codex CLI (jev-blindspot).

Open on GitHub ↗

The source publishes no measured number.
Use
Agent guardrails and approvals
Industry
Developer tools
Form
Open-source tool
Stage
In production
Listed
2026-09-25
Found via
github
Repository
jsk4581/jev-blindspot
Stars
12
Forks
0
Last push
2026-10-07
Language
TypeScript
License
MIT
jev-blindspot screenshot
docs/band.png in the jsk4581/jev-blindspot README, MIT; shown from GitHub.

The README opens with

jev-blindspot is a prompt reviewer for Claude Code and Codex CLI that runs alongside your session and shows its reviews in a local web UI.

Each time you submit a prompt, Jev decides whether a review is needed. If so, a lightweight model provides quick feedback on what your request may have overlooked, including missing context, unstated assumptions, or the underlying problem. The review helps you decide what may need clarification.

It does not block your prompt, edit it, or add anything to the agent's context. Your prompt goes through as written, and you decide whether to act on the review.

Badge

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

Listed in Jev in production

Also used for agent guardrails and approvals