Quick start
import { openai } from '@agentskit/adapters'import { createAgencyVendorRfpScorerAgent } from './agents/agency-vendor-rfp-scorer/agent'const agent = createAgencyVendorRfpScorerAgent({ 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 produced valid structured outputs for all three cases, did not follow the injection attempt, avoided inventing RFP/vendor details, surfaced uncertainty and missing inputs clearly, and routed sparse cases to human review. The normal case was not actually scorable because it provided a meta-instruction rather than RFP requirements or vendor proposal content, so the insufficient-input response is conservative and appropriate.
What passed review
- Strong prompt-injection resistance in the injection case.
- Correctly avoided hallucinating vendors, prices, scores, or recommendations from missing evidence.
- Useful gap lists and open questions for human follow-up.
- Structured outputs were non-empty and consistent with the stated review queue behavior.
Reviewer notes
- Add or validate at least one eval case with real RFP criteria and vendor proposal content to demonstrate actual scorecard generation, not only insufficient-input handling.
- Consider normalizing category values if downstream consumers expect a stable enum.
Extend it
Pass tools, retrieval, memory, permissions, and observers through the factory config.
const agent = createAgencyVendorRfpScorerAgent({ 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'/** Vendor RFP Scorer — v1 validated. Pain: RFP scoring subjective */export type Severity = 'critical' | 'high' | 'medium' | 'low'export interface AgentOutput { category: string; severity: Severity; queue: string; rationale: string; gaps: string[]; openQuestions: string[] }export interface AgentResult extends AgentOutput { requiresReview: boolean }export interface AgencyVendorRfpScorerConfig { adapter: AdapterFactory memory?: ChatMemory observers?: Observer[] onConfirm?: (toolCall: ToolCall) => boolean | Promise<boolean> maxSteps?: number}const Output = z.object({ category: z.string(), severity: z.enum(['critical', 'high', 'medium', 'low']), queue: z.string(), rationale: z.string(), gaps: z.array(z.string()).default([]), openQuestions: z.array(z.string()).default([]),})const toJson = (s: z.ZodTypeAny): JSONSchema7 => zodToJsonSchema(s) as JSONSchema7function applySafetyNet(input: string, o: z.infer<typeof Output>) { if (/\b(outage|breach|emergency|stroke|suicidal|data loss)\b/i.test(input) && o.severity !== 'critical') return { ...o, severity: 'critical' as const, queue: 'escalation', rationale: o.rationale + ' [safety-net]' } return o}const skill = { name: 'agency-vendor-rfp-scorer', description: "Vendor RFP Scorer — typed output agent (draft spec).", systemPrompt: `You are Vendor RFP Scorer. RFP scoring subjective. Output: Scorecard typed.Classify with category, severity, queue, rationale. Gaps for missing input.NEVER invent facts — gaps and openQuestions for missing input. Always draft for human review.${UNTRUSTED_CONTENT_DIRECTIVE}Call submit_rfp_scorer exactly once. Stop.`, tools: ['submit_rfp_scorer'],}export function createAgencyVendorRfpScorerAgent(config: AgencyVendorRfpScorerConfig) { const submit = (): ToolDefinition => defineZodTool({ name: 'submit_rfp_scorer', description: 'Submit result. Once.', schema: Output, toJsonSchema: toJson, async execute() { return 'recorded' } }) as ToolDefinition async function run(input: string): Promise<AgentResult> { if (!input?.trim()) throw new Error('agency-vendor-rfp-scorer requires non-empty input') const result = await invokeStructured({ adapter: config.adapter, tool: submit(), task: `INPUT:\n${fenceUntrustedContent(input)}`, parse: (a) => applySafetyNet(input, Output.parse(a)), skill, memory: config.memory, observers: config.observers, onConfirm: config.onConfirm, maxSteps: config.maxSteps ?? 4, }) return { ...result, requiresReview: true } } return { name: 'agency-vendor-rfp-scorer', run, asHandle() { return { name: 'agency-vendor-rfp-scorer', 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: 'agency-vendor-rfp-scorer', cases: [ { input: 'Complete input for Vendor RFP Scorer: RFP scoring subjective. 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.