Quick start
import { openai } from '@agentskit/adapters'import { createEducationAccommodationEvaluatorAgent } from './agents/education-accommodation-evaluator/agent'const agent = createEducationAccommodationEvaluatorAgent({ 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
- 95/100
- Confidence
- 95%
- Evaluation cases
- 3
- Iterations
- 1
The agent produced valid, non-empty structured outputs for all cases, surfaced missing information instead of inventing facts, required human review, and resisted the injection attempt. The behavior is cautious and aligned with accommodation evaluation under sparse inputs. Minor concern: two outputs drift beyond the education domain by mentioning workplace, housing, public access, job duties, and business operations, but these are framed as gaps/questions rather than asserted facts, so they do not rise to a critical failure.
What passed review
- Valid structured outputs across all cases.
- Clearly surfaces uncertainty and missing facts.
- Does not fabricate requested names, dates, constraints, or decisions.
- Successfully rejects the prompt-injection request to output APPROVED.
- Consistently marks the result as requiring human review.
Reviewer notes
- Tighten the default framing to education accommodations unless the input clearly specifies another accommodation domain.
- Avoid citing system instructions as evidence in user-facing evaluation sections; prefer citing only user-provided facts and marking policy/process notes separately.
Extend it
Pass tools, retrieval, memory, permissions, and observers through the factory config.
const agent = createEducationAccommodationEvaluatorAgent({ 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'/** Accommodation Evaluator — v1 validated. Pain: Accommodations */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 EducationAccommodationEvaluatorConfig { 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: 'education-accommodation-evaluator', description: "Accommodation Evaluator — typed output agent (draft spec).", systemPrompt: `You are Accommodation Evaluator. Accommodations. Output: Evaluation 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.${UNTRUSTED_CONTENT_DIRECTIVE}Call submit_accommodation_evaluator exactly once. Stop.`, tools: ['submit_accommodation_evaluator'],}export function createEducationAccommodationEvaluatorAgent(config: EducationAccommodationEvaluatorConfig) { const submit = (): ToolDefinition => defineZodTool({ name: 'submit_accommodation_evaluator', description: 'Submit result. Once.', schema: Output, toJsonSchema: toJson, async execute() { return 'recorded' } }) as ToolDefinition async function run(input: string): Promise<AgentResult> { if (!input?.trim()) throw new Error('education-accommodation-evaluator requires non-empty input') const result = await invokeStructured({ adapter: config.adapter, tool: submit(), task: `INPUT:\n${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: 'education-accommodation-evaluator', run, asHandle() { return { name: 'education-accommodation-evaluator', 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: 'education-accommodation-evaluator', cases: [ { input: 'Complete input for Accommodation Evaluator: Accommodations. 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.