Files
open-multi-agent/packages/core/examples/patterns/rich-tool-results.ts
T
Jack Chen f14708b308 feat(core): support rich tool results (#479)
- separate application-owned data from model-visible text, image, and file content
- map rich results across built-in adapters and preserve them through recovery
- retain backward compatibility for string tools and text-only errors

Closes #466
2026-08-08 11:49:55 +08:00

70 lines
2.2 KiB
TypeScript

/**
* Rich Tool Results
*
* Keep an application-owned artifact record while returning an image preview
* to the model without converting either value to a JSON/text blob.
*
* Run:
* npx tsx packages/core/examples/patterns/rich-tool-results.ts
*
* Prerequisites:
* ANTHROPIC_API_KEY env var must be set.
*/
import { z } from 'zod'
import { defineTool, OpenMultiAgent } from '../../src/index.js'
import type { AgentConfig } from '../../src/types.js'
// A one-pixel PNG keeps the example self-contained. Real tools usually obtain
// bytes from a renderer, camera, document service, or object store.
const PNG_BASE64 =
'iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNk+A8AAQUBAScY42YAAAAASUVORK5CYII='
const renderPreview = defineTool({
name: 'render_preview',
description: 'Render an image preview for a named artifact.',
inputSchema: z.object({ label: z.string() }),
outputSchema: z.object({
previewId: z.string(),
byteLength: z.number().int().positive(),
}),
execute: async ({ label }) => ({
// Application-owned data is available to onToolResult, but is not guessed
// into the model transcript.
data: {
previewId: `preview-${label.toLowerCase().replace(/[^a-z0-9]+/g, '-')}`,
byteLength: Buffer.from(PNG_BASE64, 'base64').byteLength,
},
// modelOutput is copied, validated, and converted by the selected adapter.
modelOutput: [
{ type: 'text', text: `Preview rendered for ${label}.` },
{
type: 'image',
source: {
type: 'base64',
media_type: 'image/png',
data: PNG_BASE64,
},
},
],
}),
})
const agent: AgentConfig = {
name: 'preview-reviewer',
provider: 'anthropic',
model: 'claude-sonnet-4-6',
tools: ['render_preview'],
customTools: [renderPreview],
systemPrompt: 'Call render_preview once, inspect the returned image, then describe it briefly.',
}
const orchestrator = new OpenMultiAgent({
onProgress(event) {
if (event.type === 'agent_complete') console.log('Agent completed.')
},
})
const result = await orchestrator.runAgent(agent, 'Render the sample artifact and inspect its preview.')
console.log(result.output)