评估与模拟

const result = await client.agents.evaluate("agent-id", {
templateId: "template-id",
messages: [
  { role: "user", content: "I'm feeling really stressed about work" },
  { role: "assistant", content: "I hear you. Work stress can be overwhelming..." },
],
});

console.log(result.score);       // 0-100
console.log(result.feedback);    // detailed feedback
console.log(result.categories);  // per-category scores

// Create a template
const template = await client.evalTemplates.create({
name: "Empathy & Support",
description: "Evaluates emotional intelligence and supportive responses",
scoringRubric: "Score based on empathy, active listening, and actionable advice",
categories: ["empathy", "active_listening", "actionable_advice"],
judgeModel: "claude-sonnet-4-6",
temperature: 0.3,
});

// List templates
const templates = await client.evalTemplates.list();

for await (const event of client.agents.simulate("agent-id", {
maxSessions: 3,
maxTurnsPerSession: 10,
simulatedDurationHours: 24,
enableProactive: true,
enableConsolidation: true,
userPersonas: [
  {
    name: "Alex",
    background: "College student struggling with math",
    personalityTraits: ["anxious", "eager to learn"],
    communicationStyle: "casual, uses slang",
  },
],
})) {
console.log(`[${event.type}] ${event.message}`);
if (event.totalCostUsd) {
  console.log(`Cost so far: $${event.totalCostUsd}`);
}
}

for await (const event of client.agents.runEval("agent-id", {
templateId: "template-id",
maxSessions: 5,
maxTurnsPerSession: 8,
})) {
if (event.type === "evaluation") {
  console.log("Score:", event.score);
}
}

// List runs
const runs = await client.evalRuns.list({ agentId: "agent-id" });

// Get a specific run
const run = await client.evalRuns.get("run-id");

// Reconnect to a streaming run
for await (const event of client.evalRuns.streamEvents("run-id")) {
console.log(event.type, event.message);
}