@mindburn/helm-openai-agents

HELM governance adapter for the OpenAI Agents SDK.

What it does

Wraps tool definitions with HELM governance so every tool call:

1. Is evaluated against HELM policy before execution
2. Is denied if the policy rejects it (fail-closed by default)
3. Produces a receipt for the execution proof chain

Quick start

import { HelmToolProxy } from "@mindburn/helm-openai-agents";

const proxy = new HelmToolProxy({
  baseUrl: "http://localhost:8080",
  apiKey: process.env.HELM_API_KEY,
});

// Wrap your tools
const governedTools = proxy.wrapTools(myTools);

// Use in an agent — every call now goes through HELM
const agent = new Agent({ tools: governedTools });

Configuration

| Option | Default | Description | | ----------------- | -------- | -------------------- | | baseUrl | required | HELM kernel URL | | apiKey | optional | HELM API key | | failClosed | true | Deny on HELM errors | | collectReceipts | true | Keep receipt chain | | onReceipt | — | Callback per receipt | | onDeny | — | Callback per denial |

License

Apache-2.0