Data Shaping & Validation: Trusted Records, Safe Coercion & Quarantine
outcome
Turn whatever a source actually sends into a shape the rest of the flow can trust, and quarantine what it cannot — without silently rewriting real values.
units
- U1
Validate at the boundary
schema first, before anything costs money
- U2
Falsy values & coercion
?? vs ||, string numbers, string booleans
- U3
Arrays & nested reads
safe collection handling, empty is not an error
- U4
The quarantine branch
set aside, reconcile, alert on a spike
- U5
Canonical forms
one spelling per concept, normalised on the way in
sandbox validation — how you pass
Build the flow in your engine and run it. When it works, export the flow JSON and send this payload to your unique validation URL. Headers and structure are verified server-side; the module flips to COMPLETED only on a correct hit.
POST https://learnflowlogic.com/api/sandbox/{your-webhook-id}
X-FlowLogic-Signature: hmac-sha256=<hex digest of the raw body>
{
"module": "data-shaping-and-validation",
"artifact": {
"flow_id": "<flow id from your engine URL>",
"run_evidence": { "status": 200, "attempts": 1 }
}
}Sign in to get your unique validation URL and signing secret.
zero-cost sandbox // data shaping & validation: trusted recor