What it does
The gate acts when the chosen category’s probability reaches 0.9. Thank-you messages receive a fixed Hebrew reply; mistakes can suppress a legitimate turn.
Agent tooling
jev-triage
Let Jev sort incoming OpenClaw messages before the main model runs, replying to thanks and silently dropping confident spam judgments.
Bookmark: jev-triage
Keep this in your collection.
Leave a noteWhat would you try with this?
: jev-triage
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
How you can use it
You can use this add-on to sort incoming chat messages before your main chatbot answers them. Your developer connects it to OpenClaw using an account access key from TypeSafe, an outside AI service. The tool inspects each message and handles simple cases directly, saving your main bot for real questions.
Have your developer test sample messages first. The filter only acts on its own when it is very confident. It silently ignores suspected spam and sends a preset Hebrew thank-you reply to grateful users. If it guesses wrong, it might accidentally silence a real customer question.