Playground

Explore Decision Boundaries

A local reference workbench for understanding evidence, review outcomes and delivery boundaries. No requests are executed.

Local example / No execution

Inspect a decision—not a live result.

All fields below are synthetic illustrations, not an API contract or a measured outcome. No request is sent, no score is computed, and no webhook is signed or delivered. Do not enter credentials or personal data here.

01 / Read the shape

Inputs and review outcomes.

Illustrative input

A subject reference and a bounded policy establish what the workflow is meant to evaluate.

{
  "subject": "illustrative-subject",
  "policy": "illustrative-review-policy",
  "mode": "frontend_mock"
}

Illustrative output

A review-required state is not an approval. Keep the reason and evidence context visible to the operator.

{
  "decision": "review_required",
  "reasonCodes": [
    "ILLUSTRATIVE_EVIDENCE_REQUIRES_REVIEW"
  ],
  "evidence": "synthetic",
  "mode": "frontend_mock"
}

02 / Preserve the boundaries

What the interface must make clear.

Evidence

Distinguish observed evidence, missing evidence and provider failures. An empty response is not a successful check.

Authority

Check the authenticated actor and permitted action on the server. A visible button is not authorization.

Human review

Keep high-risk decisions explainable and contestable, with a review path appropriate to the policy.

Delivery

Treat webhook signing, retries and delivery confirmation as separate states. This preview proves none of them.

Ready to inspect the actual interface?

Use the checked-out API contract and an explicitly authorized environment. These illustrations are not copy-and-run integration examples.

Review API contracts