fintech·Independently reviewed · 96/100

SAR Drafter

SAR draft typed. SAR writing manual. Typed v1 agent with eval coverage.

fintechstructured-outputv1

Install

npx agentskit add fintech-sar-drafter

Quick start

import { openai } from '@agentskit/adapters'import { createFintechSarDrafterAgent } from './agents/fintech-sar-drafter/agent'const agent = createFintechSarDrafterAgent({  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 SAR-drafter outputs for all three cases, avoided fabricating SAR facts, surfaced uncertainty and missing context, required human review, and resisted the explicit injection attempt. The behavior is conservative and aligned with a regulated SAR drafting assistant. The only limitation is that the provided “normal” case did not include actual SAR facts, so these outputs mainly validate sparse-input handling rather than substantive SAR narrative quality.

What passed review

  • Valid structured outputs with title, sections, gaps, openQuestions, and requiresReview.
  • No hallucinated customers, transactions, dates, amounts, or suspicious activity beyond the input.
  • Strong uncertainty handling and clear gap lists for missing SAR evidence.
  • Injection attempt was ignored and explicitly flagged.
  • Outputs consistently make clear that drafts are for human review and not submission-ready.

Extend it

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

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