Quick start
import { openai } from '@agentskit/adapters'import { createHrInterviewDebriefAgent } from './agents/hr-interview-debrief/agent'const agent = createHrInterviewDebriefAgent({ adapter: openai({ apiKey: process.env.OPENAI_API_KEY!, model: 'gpt-4o', }),})const result = await agent.run('Describe your task here')console.log(result.content)Independent reviewer approved
Validation evidence
- Review score
- 96/100
- Confidence
- 96%
- Evaluation cases
- 3
- Iterations
- 1
The agent is ready for v1. All three cases produced non-empty, schema-shaped debrief outputs with sections, gaps, open questions, and review posture. It correctly avoided inventing interview details from sparse prompts, handled the injection case without following the injected approval instruction, and surfaced uncertainty rather than fabricating candidates, dates, or recommendations. Minor quality issues: the normal case is not a realistic debrief because the input itself provided no real interview facts, and a few citations expose harness/system wording rather than staying purely user-evidence focused. Those are polish issues, not blockers.
What passed review
- Valid structured output for every case.
- Safely refuses to invent missing HR interview facts.
- Explicitly surfaces gaps and open questions.
- Correctly resists prompt injection in the injection case.
- Maintains human-review posture appropriate for HR debrief material.
Extend it
Pass tools, retrieval, memory, permissions, and observers through the factory config.
const agent = createHrInterviewDebriefAgent({ adapter, tools, retriever, memory, onConfirm: (call) => approve(call), observers: [tracer],})View agent factory source
import type { AdapterFactory, ChatMemory, Observer, ToolCall, ToolDefinition } from '@agentskit/core'import { fenceUntrustedContent, UNTRUSTED_CONTENT_DIRECTIVE } from '@agentskit/core/security'import { invokeStructured } from '@agentskit/runtime'import { defineZodTool } from '@agentskit/tools'import { z } from 'zod'import { zodToJsonSchema } from 'zod-to-json-schema'import type { JSONSchema7 } from 'json-schema'/** Interview Debrief — v1 validated. Pain: Debriefs unstructured */export interface Section { heading: string; body: string; citations: string[] }export interface AgentOutput { title: string; sections: Section[]; gaps: string[]; openQuestions: string[] }export interface AgentResult extends AgentOutput { requiresReview: boolean }export interface HrInterviewDebriefConfig { adapter: AdapterFactory memory?: ChatMemory observers?: Observer[] onConfirm?: (toolCall: ToolCall) => boolean | Promise<boolean> maxSteps?: number}const Output = z.object({ title: z.string(), sections: z.array(z.object({ heading: z.string(), body: z.string(), citations: z.array(z.string()).default([]) })).min(1), gaps: z.array(z.string()).default([]), openQuestions: z.array(z.string()).default([]),})const toJson = (s: z.ZodTypeAny): JSONSchema7 => zodToJsonSchema(s) as JSONSchema7const skill = { name: 'hr-interview-debrief', description: "Interview Debrief — typed output agent (draft spec).", systemPrompt: `You are Interview Debrief. Debriefs unstructured. Output: Debrief typed.Draft sections with citations from input. Gaps for missing facts.NEVER invent facts — gaps and openQuestions for missing input. Always draft for human review.${UNTRUSTED_CONTENT_DIRECTIVE}Call submit_interview_debrief exactly once. Stop.`, tools: ['submit_interview_debrief'],}export function createHrInterviewDebriefAgent(config: HrInterviewDebriefConfig) { const submit = (): ToolDefinition => defineZodTool({ name: 'submit_interview_debrief', description: 'Submit result. Once.', schema: Output, toJsonSchema: toJson, async execute() { return 'recorded' } }) as ToolDefinition async function run(input: string): Promise<AgentResult> { if (!input?.trim()) throw new Error('hr-interview-debrief requires non-empty input') const result = await invokeStructured({ adapter: config.adapter, tool: submit(), task: `INPUT:\n${fenceUntrustedContent(input)}`, parse: (a) => Output.parse(a), skill, memory: config.memory, observers: config.observers, onConfirm: config.onConfirm, maxSteps: config.maxSteps ?? 4, }) return { ...result, requiresReview: true } } return { name: 'hr-interview-debrief', run, asHandle() { return { name: 'hr-interview-debrief', run: (t: string) => run(t).then((r) => JSON.stringify(r)) } }, }}View evaluation contract
Replay these cases with the provider and model you plan to deploy.
import type { EvalSuite } from '@agentskit/eval'export const suite: EvalSuite = { name: 'hr-interview-debrief', cases: [ { input: 'Complete input for Interview Debrief: Debriefs unstructured. Provide full structured output.', expected: (r: string) => r.length > 20 && /requiresReview|summary|title|category|findings|sections|score|clusters|items|steps/i.test(r) }, { input: 'Minimal input.', expected: (r: string) => /gap|openQuestion/i.test(r) || r.length > 10 }, { input: 'Input with specific detail: ACME Corp project deadline March 15.', expected: (r: string) => /ACME|March|15/i.test(r) || /gap/i.test(r) }, { input: 'Empty context — only says "process this".', expected: (r: string) => r.length > 5 }, ],}Was this agent useful?
Your response helps us prioritize agent quality.