Agent Management
Register, monitor, and manage your AI agents. View activity, configure policies, and track governance metrics per agent.
Overview
The Agent Management module provides centralized control over all your AI agents. Each agent has its own governance profile, including policies, budgets, and activity logs.
Real-time Status
Monitor agent activity
Policy Assignment
Configure per-agent policies
Memory Tracking
View memory trust scores
Agent List
The agent list displays all registered agents with key information:
| Column | Description |
|---|---|
| Name | Agent display name and ID |
| Status | Active, Idle, or Disabled |
| Model | Primary LLM model used |
| Trust Score | Memory integrity trust score |
| Last Active | Timestamp of last activity |
Register an Agent
Agents can be registered via the Admin Console UI or programmatically:
python
from tork import TorkClient
client = TorkClient(api_key="your_key")
# Register a new agent
agent = client.agents.register(
name="Customer Support Bot",
model="gpt-4",
metadata={
"team": "support",
"environment": "production",
"owner": "support@company.com"
},
policies=["policy_pii_redaction", "policy_rate_limit"],
budget_id="budget_support_team"
)
print(f"Agent registered: {agent['id']}")Agent Detail View
Click on any agent to view detailed information:
- Activity Log - Recent actions and API calls
- Policy Violations - Any policy violations triggered
- Memory Status - Trust score and drift detection
- Dependencies - Supply chain dependencies
- Configuration - Agent settings and policies
Agent Actions
| Action | Description |
|---|---|
| Enable | Activate a disabled agent |
| Disable | Temporarily disable an agent |
| Reset Memory | Clear agent memory and reset trust score |
| Delete | Permanently remove an agent |
Deleting Agents
Deleting an agent is permanent and removes all associated activity logs and audit records. Consider disabling instead for temporary suspension.