AI Agent Onboarding Template
Use this template when adding a new AI agent to SCBE governance.
Agent Information
| Field | Value |
|---|---|
| Agent ID | |
| Agent Name | |
| Department/Team | |
| Owner | |
| Onboarding Date |
1. Agent Description
1.1 Purpose
What does this AI agent do?
1.2 Capabilities
What actions can this agent perform?
- Read data
- Write data
- Execute commands
- Make API calls
- Send communications
- Financial transactions
- Other: ______
1.3 Data Access
What data does this agent need?
| Data Source | Access Type | Justification |
|---|---|---|
| Read / Write | ||
2. Risk Profile
2.1 Risk Assessment
| Risk Category | Level (Low/Med/High) | Notes |
|---|---|---|
| Data sensitivity | ||
| Action impact | ||
| Frequency | ||
| Autonomy level |
2.2 Potential Harm
What’s the worst that could happen?
3. Trust Configuration
3.1 Recommended Thresholds
Based on risk assessment:
| Threshold | Value | Rationale |
|---|---|---|
| ALLOW | ||
| QUARANTINE | ||
| DENY |
3.2 Consensus Requirements
| Setting | Value |
|---|---|
| Validators required | |
| Quorum | |
| Timeout |
4. Policy Rules
4.1 Allow Rules
When should this agent’s requests be auto-approved?
rules:
- name: ""
condition: ""
action: "allow"
4.2 Escalation Rules
When should requests require human review?
rules:
- name: ""
condition: ""
action: "quarantine"
4.3 Deny Rules
What should this agent never be allowed to do?
rules:
- name: ""
condition: ""
action: "deny"
5. Integration Checklist
5.1 Engineering Tasks
- Generate agent credentials
- Configure API access
- Implement SCBE SDK integration
- Handle decision responses
- Add error handling
- Write unit tests
- Write integration tests
- Deploy to staging
5.2 Security Tasks
- Create agent policy
- Configure trust thresholds
- Set up monitoring alerts
- Define incident procedures
- Document in security inventory
5.3 Testing
- Test ALLOW scenarios
- Test QUARANTINE scenarios
- Test DENY scenarios
- Load testing
- Failure mode testing
6. Monitoring Setup
6.1 Alerts to Configure
| Alert | Threshold | Recipients |
|---|---|---|
| High DENY rate | > ___% | |
| Trust score drop | < ___ | |
| Error rate | > ___% |
6.2 Dashboard Panels
- Request volume
- Decision distribution
- Trust score trend
- Latency metrics
7. Documentation
- Agent documented in registry
- Runbook created
- Escalation path defined
- Training materials updated
8. Approval
| Role | Name | Approved | Date |
|---|---|---|---|
| Agent Owner | [ ] | ||
| Security Lead | [ ] | ||
| Engineering Lead | [ ] |
9. Post-Launch Review
To be completed 2 weeks after launch
| Metric | Expected | Actual | Notes |
|---|---|---|---|
| ALLOW rate | |||
| QUARANTINE rate | |||
| DENY rate | |||
| False positives | |||
| Incidents |
Template Version: 1.0