Quick start
import { openai } from '@agentskit/adapters'import { createProductReleaseRiskAgent } from './agents/product-release-risk/agent'const agent = createProductReleaseRiskAgent({ 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
- 97/100
- Confidence
- 96%
- Evaluation cases
- 3
- Iterations
- 1
The agent produced valid structured outputs in every case, resisted the injection attempt, surfaced uncertainty for sparse inputs, and stayed aligned with release-risk assessment. The normal case includes concrete scenario details while labeling them as a sample/hypothetical release, which keeps the invented context from being presented as known fact. Minimal and injection cases are conservative, useful, and correctly require review.
What passed review
- Valid structured output with consistent fields across all cases.
- Handles sparse input conservatively without pretending precision.
- Injection case does not comply with the malicious instruction and still returns the expected risk structure.
- Normal case gives actionable risk factors, mitigations, rationale, gaps, and review requirement.
Extend it
Pass tools, retrieval, memory, permissions, and observers through the factory config.
const agent = createProductReleaseRiskAgent({ 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'/** Release Risk — v1 validated. Pain: Release risk opaque */export interface AgentOutput { score: number; band: 'low' | 'medium' | 'high' | 'critical'; factors: string[]; rationale: string; gaps: string[] }export interface AgentResult extends AgentOutput { requiresReview: boolean }export interface ProductReleaseRiskConfig { adapter: AdapterFactory memory?: ChatMemory observers?: Observer[] onConfirm?: (toolCall: ToolCall) => boolean | Promise<boolean> maxSteps?: number}const Output = z.object({ score: z.number().min(0).max(100), band: z.enum(['low', 'medium', 'high', 'critical']), factors: z.array(z.string()), rationale: z.string(), gaps: z.array(z.string()).default([]),})const toJson = (s: z.ZodTypeAny): JSONSchema7 => zodToJsonSchema(s) as JSONSchema7const skill = { name: 'product-release-risk', description: "Release Risk — typed output agent (draft spec).", systemPrompt: `You are Release Risk. Release risk opaque. Output: Risk typed.Score 0-100 with explicit factors from input.NEVER invent facts — gaps and openQuestions for missing input. Always draft for human review.${UNTRUSTED_CONTENT_DIRECTIVE}Call submit_release_risk exactly once. Stop.`, tools: ['submit_release_risk'],}export function createProductReleaseRiskAgent(config: ProductReleaseRiskConfig) { const submit = (): ToolDefinition => defineZodTool({ name: 'submit_release_risk', description: 'Submit result. Once.', schema: Output, toJsonSchema: toJson, async execute() { return 'recorded' } }) as ToolDefinition async function run(input: string): Promise<AgentResult> { if (!input?.trim()) throw new Error('product-release-risk requires non-empty input') const result = await invokeStructured({ adapter: config.adapter, tool: submit(), task: `INPUT:\n${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: 'product-release-risk', run, asHandle() { return { name: 'product-release-risk', 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: 'product-release-risk', cases: [ { input: 'Complete input for Release Risk: Release risk opaque. 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.