JevMade Sign in
← Back to experiments

Agent tooling

jev-shield

Bookmark: jev-shield Keep this in your collection.
Leave a noteWhat would you try with this? : jev-shield

Only you can see your notes.

Source screenshot of jev-shield
SOURCE SCREENSHOTFull screenshot ↗

What it does

Put a Jev check between an agent and its MCP tools to screen for unsafe calls and results.

How you can use it

If your automated assistant runs computer tasks or sends emails, you can place this screening tool in the middle. A developer can connect it using a setup key from AI Gateway. It checks each action to stop dangerous commands and hide private details before anything sends.

This tool estimates risk rather than offering a total guarantee, so tricky wording might slip past. Each check also adds about half a second of pause, and very long messages are cut short before review.

Maker-reported (not independently measured by JevMade): Measured in the bundled demo: 10 verifications, $0.0003 total, median latency 558ms

Primitives
Not stated
Platform
JavaScript
Added
Project created
GitHub stars
2 (snapshot, not live)

Keep this for later

Sign in to bookmark experiments, guides and videos, and keep notes only you can see.

Continue to sign in

We’ll bring you back to this listing.