What it is
A single-tool MCP server that screens text for spam and toxicity. It returns a structured verdict with probabilities, category, severity, confidence, and a recommended action, for host agents that need a moderation gate.
How it uses Jev
Jev answers typed questions over the input text: spam and toxic probabilities, a category choice, and a severity score. The server combines these into violates_policy and an allow/review/block action, gated on confidence rather than spam_prob alone.
Primitives:choicescorenoul
Technique worth stealing
Gate auto-actions on calibrated confidence, not on a single probability threshold.
Try it
npm install && npm run build; run node dist/index.js, or doctor for diagnostics.
Evidence
Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.
- Jev-centricyes0.95
- Shows a System One patternunclear0.41
- Handles uncertaintyno0.11
- Measuredyes0.95
- Runnableyes0.95
- Worth recommendingno0.30
- Model replicano0.10
- Problem scopescore on a 0–2 scale1.84
- About Jevyes0.98
Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 23 Sept 2026.