RedHub AIRedHub AI
Tool Gate · How it works

How the roll-up gate works

How the Test-Mode / Live-Mode Line-Up gate works

This tool grades each service on whether your development and production wires are genuinely separate. It reads 10 services and rolls them into one verdict — the worst, never the average — then a gate that can only make it worse.

The 10 services

01S-01 · Your databasedev → test · prod → live · separate accountsSEPARATED
02S-02 · Paymentstwo modes, but one account behind bothDEV ON LIVE
03S-03 · Sending emaildev points straight at liveDEV ON LIVE
04S-04 · File and image storageproduction is still pointed at the sandboxPROD ON TEST
05S-05 · Your domain and DNSno test mode exists to get wrongNO SEPARATE MODES
06S-06 · Sign-in redirect addressesdev → test · prod → live · separate accountsSEPARATED
07S-07 · The AI provider your app callsfix firstdev target not foundDEV ON LIVE
08S-08 · Website analyticsno test mode exists to get wrongNO SEPARATE MODES
09S-09 · Background jobs and queuesshared-account answer could not be foundDEV ON LIVE
10S-10 · Text messages and pushproduction is still pointed at the sandboxPROD ON TEST
LINES SEPARATE
every service with two modes keeps them apart
POINTING AT THE SANDBOX
production is still wired to a test target
WIRES CROSSED
development can reach live, or one account serves both
Nominal separation · worsen-only

Two modes behind one account is DEV ON LIVE however carefully the keys are labelled. A test key and a live key that bill, send, and write to the same account are one wire with two names — the separation is nominal, and the gate treats it as no separation at all.

WIRES CROSSED

Why: Four of ten services let development reach live. Only one of them points at live outright — the rest are subtler: payments has two proper modes sitting behind a single account, and two more had an answer nobody could find. Two services are pointed the harmless way round, at a sandbox in production. Two have no test mode at all and never count against you. 4 of 10 dev-on-live · 2 prod-on-test · worst-not-average.

Fix firstS-07, the AI provider your app calls — where development points could not be found, so it resolves to live

No single service is the crisis — the clustering is. The roll-up takes the worst, the gate escalates when trouble compounds, and it names the one thing to fix first. Grades how environments are wired, never a person. Not a scanner — it reads back what you tell it, using rules you can see.