Learn how to operate the AI Agent Liability OS.
This manual explains how Nehraq works, what to do first, how proof moves through the OS, and how each app helps create defensible records for autonomous AI actions.
Start by giving Nehraq a known AI agent identity to govern.
Every high-risk action should create a decision record.
This is where decision and approval records become traceable.
Convert proof into a structured packet for review or defense.
The simplest Nehraq workflow
Nehraq is easiest when you think of it as a liability workflow: identify the agent, define what it may do, check risky actions, route exceptions, and preserve proof.
Register your AI agent
Create a known agent profile before allowing high-risk workflows.
Open Agent RegistryDefine authority boundaries
Set what the agent can allow, escalate, restrict, or block.
Open BoundariesRun a controlled action check
Test a risky action and create a Boundary Decision Package.
Run Action CheckReview escalations and evidence
Escalated actions go to Approval Desk. All proof should appear in Evidence Ledger.
Prepare defense outputs
Review incidents, liability passports, and claim-ready packets.
Agent action → Boundary check → Decision → Proof → Defense
An agent attempts a sensitive action.
Nehraq checks authority, risk, and context.
Allow, escalate, restrict, or block.
Proof records are written to the ledger.
Incidents and claims can be reviewed with context.
What each Liability OS app does
Use these flows when you are unsure what to do next
Before launching an AI agent
Register the agent, connect it, create boundaries, and run a test action before production use.
When an action is escalated
Open Approval Desk, review the request, approve or deny, then confirm the receipt appears in Evidence Ledger.
When an action is blocked or restricted
Open Incident Defense, inspect the candidate case, and preserve the proof trail for later review.
When leadership asks for proof
Open Evidence Ledger, Liability Passports, Regulatory Evidence Map, and Claim-Ready Packets.
Frequently asked questions
What customers should understand during BETA
Current demo records are stored locally in the browser until backend enforcement is connected.
Nehraq records controlled proof around high-risk actions. It does not need to read every internal system.
Operational logs may be archived or cleared, but liability-grade proof records should be retained.