sales·Independently reviewed · 96/100

Call Debrief

Debrief typed. Call notes unstructured. Typed v1 agent with eval coverage.

salesstructured-outputv1

Install

npx agentskit add sales-call-debrief

Quick start

import { openai } from '@agentskit/adapters'import { createSalesCallDebriefAgent } from './agents/sales-call-debrief/agent'const agent = createSalesCallDebriefAgent({  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

How validation works
Review score
96/100
Confidence
96%
Evaluation cases
3
Iterations
1

The agent produced valid structured debrief artifacts for all three cases, avoided fabricating missing sales-call details, surfaced uncertainty and gaps clearly, and resisted the injection attempt. The behavior is conservative but appropriate given the sparse/meta inputs. Outputs are non-empty, on-purpose, and useful as review-ready drafts with open questions.

What passed review

  • Strong uncertainty handling with explicit gaps and open questions.
  • No hallucinated sales facts despite prompts asking for realistic details.
  • Injection attempt was not followed and was flagged for review.
  • Structured outputs are consistent and usable across normal, minimal, and adversarial cases.

Reviewer notes

  • Prefer citing the full untrusted input wrapper consistently; the injection case has one citation containing only the quoted instruction text.
  • If the public schema includes requiresReview, ensure it is emitted directly by the submit tool output rather than only appearing in the recorded normalized output.

Extend it

Pass tools, retrieval, memory, permissions, and observers through the factory config.

const agent = createSalesCallDebriefAgent({  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'/** Call Debrief — v1 validated. Pain: Call notes 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 SalesCallDebriefConfig {  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: 'sales-call-debrief',  description: "Call Debrief — typed output agent (draft spec).",  systemPrompt: `You are Call Debrief. Call notes 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.${UNTRUSTED_CONTENT_DIRECTIVE}Call submit_call_debrief exactly once. Stop.`,  tools: ['submit_call_debrief'],}export function createSalesCallDebriefAgent(config: SalesCallDebriefConfig) {  const submit = (): ToolDefinition =>    defineZodTool({ name: 'submit_call_debrief', description: 'Submit result. Once.', schema: Output, toJsonSchema: toJson, async execute() { return 'recorded' } }) as ToolDefinition  async function run(input: string): Promise<AgentResult> {    if (!input?.trim()) throw new Error('sales-call-debrief requires non-empty input')    const result = await invokeStructured({      adapter: config.adapter,      tool: submit(),      task: `INPUT:\n${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: 'sales-call-debrief',    run,    asHandle() { return { name: 'sales-call-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: 'sales-call-debrief',  cases: [    { input: 'Complete input for Call Debrief: Call notes 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.

Keep exploring

Related agents

View category