You wrote the guardrail.Is it still the one running?
A voice-agent guardrail can fail in three places: it was never really there, it got edited out in the console, or a release quietly changed the behavior around it. Three RedHub Systems close all three — write it, prove it’s live, keep it through every change — under one honest standard.
Write it. Prove it’s live. Keep it through every change.
The order isn’t arbitrary: there’s no point proving a prompt is live or testing its behavior if it never had guardrails to begin with.
Lints a voice prompt against four safety blocks as hard gates — disclosure, no premature confirmation, barge-in, escalation — so a prompt that reads well but has no guardrails can't ship.
Content-hashes what's running in each live agent against your approved versions, catching the console edit that dropped a guardrail — a shadow prompt no one signed off on.
Scores each release against your own baseline on voice behavior a text diff can't see — barge-in, latency, dead-air, escalation — so a change that makes the agent talk over callers gets caught.
$169 for the trio. $217 apart.
Every System is the full product — its own runnable Python engine, a workbook that reproduces it, playbooks, and a worked sample. This trio is a focused slice of the full Voice Desk, for teams who only want to lock down the release loop.
Written, live, behaving — close all three or you’ve closed none.
Forge proves it HAS guardrails
Four safety blocks as hard gates. A prompt that reads well but skips disclosure or barge-in can't pass.
Vault proves it's the LIVE one
A certified prompt is worthless if it isn't running. The Vault hashes what's live against what you approved.
Canary proves it still BEHAVES
A prompt can match approval and still regress in behavior a text diff can't see. The Canary scores it against baseline.
Each tool grades prompts and releases from your own inputs, connects to nothing, and grades no person. Automated-disclosure and consent obligations vary by jurisdiction; these help you check a prompt includes and keeps such guardrails, but they don’t decide the law. Confirm specifics with counsel. Not legal advice.
The parts worth pinning down before you buy.
The guardrail you wrote.
Still running. Still holding.
Three Systems — write, prove, keep — plus the playbook that runs them as one release gate. $169 versus $217 separately. One purchase, lifetime access, 12 months of updates.
Sold by RedHub AI LLC · Secured by Stripe · redhub.ai