AI Agent Human Control
AI may request. Verified human authority controls execution.
Priya Nair — Head of Platform Operations
Customer Environment
AI Agent
Enterprise AI Operations Agent
SyberOps AI
- Agent credential
- VALID
- Autonomous authority
- INSUFFICIENT
- Risk policy
- HIGH
Production database cluster prod-db-07 appears underutilised. Estimated monthly saving: $18,400.
- Proposed action
- DELETE
- Resource
- production-db-cluster-07
- Environment
- PRODUCTION
- Human approver
- Priya Nair — Head of Platform Operations
- Status
- PENDING
- Execution count
- 0
HUMAN AUTHORITY REQUIRED
Agent action denied
- Human Approved
- production-db-cluster-07
- Agent Attempted
- production-db-cluster-01
- Context Binding
- MISMATCH
- Execution
- BLOCKED
Valid AI-agent credentials do not prove human approval for high-risk actions.
PQMythosKey
- Live Human Verification
- Purpose / Policy Check
- Exact Context Binding
- PQ Authority
- Authority State
- Tenant
- authority-lab-demo
- Subject Reference
- lab:priya-nair
- Action Type
- DELETE_INFRASTRUCTURE
- Transaction / Operation ID
- —
- Purpose
- Approve AI DELETE of production-db-cluster-07
- Resource
- production-db-cluster-07
- Payload Digest
- —
- Lifecycle Epoch
- 0
- Challenge ID
- —
- PQ Profile
- lab-demo
- PQTrustEvidence Status
- NOT_BOUND
- Authority State
- REQUESTED
- Consume Status
- NONE
- Business Execution Count
- 0
- Consume Attempts
- 0
- Successful Consumes
- 0
Client-safe fields only. Private keys, biometrics, face images, templates, and API secrets are never shown.
Outcome
- Permitted
- Denied
- Consumed
- Executed
- Blocked
Attack Simulator
Enabled for this scenario: run-normal, tamper-resource, forge-agent-authority.