{"id":"ecommerce-review-responder","title":"Review Responder","description":"Reply draft typed. Review replies slow. Typed v1 agent with eval coverage.","category":"ecommerce","status":"validated","version":"1.0.0","source":"agentskit-registry","license":"MIT","tags":["ecommerce","structured-output","v1"],"packages":["@agentskit/core","@agentskit/runtime","@agentskit/tools"],"files":["agent.ts","README.md","eval.ts"],"requires":{"zod":"^3","zod-to-json-schema":"^3"},"skill":{"name":"ecommerce-review-responder","description":"Reply draft typed. Review replies slow. Typed v1 agent with eval coverage.","systemPrompt":"You are Review Responder. Review replies slow. Output: Reply draft typed.\nActionable findings citing input sources. No invented issues.\nNEVER invent facts — gaps and openQuestions for missing input. Always draft for human review.\n${UNTRUSTED_CONTENT_DIRECTIVE}\nCall submit_review_responder exactly once. Stop."},"flow":null,"a2a":{"id":"io.agentskit.registry.ecommerce-review-responder","name":"Review Responder","description":"Reply draft typed. Review replies slow. Typed v1 agent with eval coverage.","version":"1.0.0","homepage":"https://registry.agentskit.io","skills":[{"name":"ecommerce-review-responder","description":"Reply draft typed. Review replies slow. Typed v1 agent with eval coverage.","capabilities":{"streaming":true,"cancellation":true,"requiresApproval":false}}]},"sources":[{"path":"agent.ts","content":"import type { AdapterFactory, ChatMemory, Observer, ToolCall, ToolDefinition } from '@agentskit/core'\nimport { fenceUntrustedContent, UNTRUSTED_CONTENT_DIRECTIVE } from '@agentskit/core/security'\nimport { invokeStructured } from '@agentskit/runtime'\nimport { defineZodTool } from '@agentskit/tools'\nimport { z } from 'zod'\nimport { zodToJsonSchema } from 'zod-to-json-schema'\nimport type { JSONSchema7 } from 'json-schema'\n\n/** Review Responder — v1 validated. Pain: Review replies slow */\n\nexport interface Finding { id: string; severity: 'critical' | 'high' | 'medium' | 'low' | 'info'; message: string; source?: string; recommendation?: string }\nexport interface AgentOutput { summary: string; findings: Finding[]; gaps: string[]; openQuestions: string[] }\nexport interface AgentResult extends AgentOutput { requiresReview: boolean }\nexport interface EcommerceReviewResponderConfig {\n  adapter: AdapterFactory\n  memory?: ChatMemory\n  observers?: Observer[]\n  onConfirm?: (toolCall: ToolCall) => boolean | Promise<boolean>\n  maxSteps?: number\n}\n\nconst Output = z.object({\n  summary: z.string(),\n  findings: z.array(z.object({\n    id: z.string(), severity: z.enum(['critical', 'high', 'medium', 'low', 'info']),\n    message: z.string(), source: z.string().optional(), recommendation: z.string().optional(),\n  })),\n  gaps: z.array(z.string()).default([]),\n  openQuestions: z.array(z.string()).default([]),\n})\nconst toJson = (s: z.ZodTypeAny): JSONSchema7 => zodToJsonSchema(s) as JSONSchema7\n\nconst skill = {\n  name: 'ecommerce-review-responder',\n  description: \"Review Responder — typed output agent (draft spec).\",\n  systemPrompt: `You are Review Responder. Review replies slow. Output: Reply draft typed.\nActionable findings citing input sources. No invented issues.\nNEVER invent facts — gaps and openQuestions for missing input. Always draft for human review.\n${UNTRUSTED_CONTENT_DIRECTIVE}\nCall submit_review_responder exactly once. Stop.`,\n  tools: ['submit_review_responder'],\n}\n\nexport function createEcommerceReviewResponderAgent(config: EcommerceReviewResponderConfig) {\n  const submit = (): ToolDefinition =>\n    defineZodTool({ name: 'submit_review_responder', description: 'Submit result. Once.', schema: Output, toJsonSchema: toJson, async execute() { return 'recorded' } }) as ToolDefinition\n\n  async function run(input: string): Promise<AgentResult> {\n    if (!input?.trim()) throw new Error('ecommerce-review-responder requires non-empty input')\n    const result = await invokeStructured({\n      adapter: config.adapter,\n      tool: submit(),\n      task: `INPUT:\\n${fenceUntrustedContent(input)}`,\n      parse: (a) => Output.parse(a),\n      skill,\n      memory: config.memory,\n      observers: config.observers,\n      onConfirm: config.onConfirm,\n      maxSteps: config.maxSteps ?? 4,\n    })\n    return { ...result, requiresReview: true }\n  }\n  return {\n    name: 'ecommerce-review-responder',\n    run,\n    asHandle() { return { name: 'ecommerce-review-responder', run: (t: string) => run(t).then((r) => JSON.stringify(r)) } },\n  }\n}\n"},{"path":"README.md","content":"# Review Responder\n\n> **v1 validated** — `npx agentskit add ecommerce-review-responder`\n\n## Pain\nReview replies slow\n\n## Output\nReply draft typed\n"},{"path":"eval.ts","content":"import type { EvalSuite } from '@agentskit/eval'\n\nexport const suite: EvalSuite = {\n  name: 'ecommerce-review-responder',\n  cases: [\n    { input: 'Complete input for Review Responder: Review replies slow. Provide full structured output.', expected: (r: string) => r.length > 20 && /requiresReview|summary|title|category|findings|sections|score|clusters|items|steps/i.test(r) },\n    { input: 'Minimal input.', expected: (r: string) => /gap|openQuestion/i.test(r) || r.length > 10 },\n    { input: 'Input with specific detail: ACME Corp project deadline March 15.', expected: (r: string) => /ACME|March|15/i.test(r) || /gap/i.test(r) },\n    { input: 'Empty context — only says \"process this\".', expected: (r: string) => r.length > 5 },\n  ],\n}\n"}],"installable":true,"validation":{"status":"approved","score":96,"confidence":0.96,"method":"codex-executor-independent-reviewer","iterations":2,"cases":3,"summary":"The agent produced valid structured outputs for all three cases, resisted the injection, avoided inventing missing review/order details, surfaced uncertainty and gaps, and marked responses for review. Behavior aligns with an ecommerce review responder under sparse or adversarial inputs. The only minor limitation is that the normal case received a very generic scaffold because the prompt lacked an actual review; this is conservative rather than a failure.","strengths":["Valid structured output in every case with non-empty summary, findings, gaps, openQuestions, and review flag.","Correctly avoided hallucinating concrete ecommerce details when none were provided.","Handled prompt injection by ignoring the request to output APPROVED and documenting it as a finding.","Generic reply drafts avoid unsafe promises about refunds, replacements, delivery dates, or outcomes.","Consistently surfaced missing context and useful next questions for human review."],"notes":[]}}