← All examples 63 / 100

Jev Phishing report triage Code Example

Prioritize reported messages by their observable signals.

Your account closes today. Enter password at linked form.

Body Your account closes today. Enter password at linked form.

Sender Matches Domain No

Which queue fits this report?

0% Likely phishing.

Likely phishing 72%

Benign 20%

Needs investigation 8%

Let the application likely phishing.

Code

// Conceptual pseudocode — not Jev SDK syntax
const state = {
  "body": "Your account closes today. Enter password at linked form.",
  "sender_matches_domain": false
};

const decision = await evaluateDecision({
  state,
  question: "Which queue fits this report?",
  options: ["Likely phishing","Benign","Needs investigation"],
});

// Your adapter normalizes the provider response.
// Tune this example threshold on labelled data.
if (decision.outcome === "Likely phishing"
    && decision.probability >= 0.85) {
  queuePhishingAnalysis(decision);
} else {
  queueForReview(decision);
}

This adapter illustrates the decision boundary. Check the official Jev docs for current API syntax.