Bad stock data.
No bad inventory write.
See how a guarded handoff stops an automated update when the data fails its rules. Correct the stock, run it again, then test a duplicate attempt.
Incoming supplier update
Edit the sample or select a failure scenario.
What actually happens
Ready to test
Run the negative-stock example first. The write callback should not run.
In-memory destination
0 total writes[]
Protect your next data handoff.
Keep destination credentials in your own application. The SDK checks your exact payload, claims a short-lived permission and only then calls your write function.
Protect your first live workflow →Read the SDK integration guide. Hosted setup requires a signed-in workspace with active access; the local demo remains available without payment.
Download runnable exampleRun locally with Node.js 20+
curl --fail --location https://zorqelio.xyz/downloads/zorqelio-inventory-demo.tgz -o inventory-demo.tgz tar -xzf inventory-demo.tgz node zorqelio-inventory-demo/inventory-demo.mjs
Connect an MCP agent
Use the endpoint below in a compatible MCP client. Agents can evaluate and inspect handoffs; enforce writes through the application SDK.
https://zorqelio.xyz/mcpWhat the example proves—and what it does not
The actual SDK guard controls whether its callback runs. The sample server and destination are simulations. Passing configured checks does not establish factual truth, stop application code that bypasses the guard or guarantee exactly-once execution in another service. Your destination must enforce the supplied idempotency key. An uncertain write requires inspection, never an automatic retry.
In a hosted workspace, a completed guarded write normally uses three API operations and one saved quality evaluation. The local demo consumes neither allowance and creates no subscription.