Skip to main content

​
Fraud Shield — Tenant Triage & Chargeback Evidence

Orbit’s detectors continuously screen your account for fraud and abuse signals — SIM-swap, IRSF / toll-fraud, SMS pumping / artificial traffic inflation (AIT), Wangiri one-ring callbacks, spend-velocity anomalies, API-key geo-anomalies, voice-deepfake scores, and suspicious campaign send velocity. Until recently only Orbit’s operators could see those alerts. Fraud Shield puts the same ledger in your hands: you can list the alerts raised against your own account, triage each one (acknowledge or dismiss), set which severities and categories count as actionable in your view, and export a CSV evidence pack suitable for a carrier chargeback dispute. The endpoints you will use section below is the consolidated index; reads are open to any authenticated member of your organization, and writes are restricted to owners and admins.

​
Categories and severities

Categories detected by the platform: Severities are low, medium, high, and critical, ranked in that order. Review statuses are open (unattended), triaged (acknowledged), dismissed (closed by you as a false positive or accepted risk), and escalated (closed by Orbit’s operators).

​
Endpoints you will use

Everything below is rooted at https://api.orbit.devotel.io/api/v1/compliance/fraud-reviews. Reads are open to any authenticated member of your organization; writes are restricted to owners and admins. A list row carries the fields you triage on — category, severity, detected_at, and the derived actionable flag — so a script can page through the list and act without opening each alert first:
actionable is true when the row meets your notify_min_severity floor and its category is not in muted_categories — critical rows are always actionable. Filter on it client-side to drive a playbook, or open individual rows with GET /compliance/fraud-reviews/:id.

​
List and inspect alerts

GET /compliance/fraud-reviews returns your organization’s alerts, sorted critical-first and newest-first. Each row carries an actionable flag derived from your configured thresholds (see below). Any PII in the alert details is scrubbed on the way out. Query parameters:
GET /compliance/fraud-reviews/:id returns one alert. Because every read is scoped to your organization, an id that belongs to another organization returns 404 — you never see cross-tenant data, and a guessable id leaks nothing.

​
Triage: acknowledge and dismiss

Acknowledging an alert means “we have seen this and are investigating.” Dismissing it means “reviewed and closed as a false positive or an accepted risk.” Both are owner/admin-only and recorded in your audit log.
Rules to know:
  • Acknowledging an already-triaged alert is a no-op (idempotent).
  • Dismissing an already-dismissed alert is a no-op.
  • A dismissed or escalated alert cannot be acknowledged; an escalated alert cannot be dismissed either way — you get a 409 explaining why.
The internal operator id of whoever triaged the alert is not exposed through these endpoints.

​
Triage thresholds shape your view — never the alert webhook

GET and PUT /compliance/fraud-reviews/settings read and update two fields:
The important contract: these thresholds are a read/triage control only. They change which alerts are flagged actionable in your list and on the dashboard — they do not silence the account.fraud.alert webhook or the in-app notification. That fan-out stays always-on so an attacker draining your wallet can never mute the warning. critical alerts are always actionable regardless of your notify_min_severity floor. The webhook catalog entry is documented in the webhooks events reference.

​
Chargeback evidence pack (CSV)

When a vendor bills you for SMS pumping / AIT or Wangiri / toll-fraud traffic, disputing the charge takes evidence. Export the blocks Orbit raised on your account as an RFC-4180 CSV that a carrier can consume:
window_days is the look-back window (1–90, default 30). The file includes one row per eligible block — SMS pumping / AIT over messaging and Wangiri / toll-fraud over voice — with a stable set of columns (review_id, detected_at, source, category, severity, status, phone_prefix, channel, rule, block_category, risk_score, summary, report_generated_at). If no eligible blocks exist over the window, you get a header-only file. Every export is written to your audit log with the window and eligible row count, so the export itself is part of the evidence trail. Verify (OTP) traffic has its own dedicated export — GET /verify/conversion-anomaly/evidence-export.csv.

​
Dashboard

The same triage surface is available in the dashboard under Settings → Compliance → Fraud Shield: the alert list with acknowledge / dismiss actions, the threshold controls above, and the chargeback evidence download. Owners and admins see the write actions; read-only members see the list and settings.

​
A tenant-owned posture note

These controls are yours. Orbit raises the alerts and keeps the webhook fan-out always-on, but acknowledging, dismissing, and deciding which signals need a human can be tuned to your own incident-response workflow. The ledger is scoped so no tenant can ever see another tenant’s alerts, and the internal operator ids behind triage decisions are never exposed to your side.
  • Voice Destination Auto-Blocks — the automatic, self-expiring per-destination voice blocks the platform applies on a traffic-pumping or Wangiri burst; each block lands in this same Fraud Shield triage queue when it rejects a dial.
  • Compliance Posture Overview — the toggle map that lists the read-only GET /settings/compliance/voice-destination-blocks endpoint alongside every other posture control.