Agents
GoZen can act as an operations layer for coding agents such as Claude Code, Codex, and other CLI-driven assistants. It helps you coordinate agent work, monitor sessions, and apply runtime safety controls without changing your existing workflow.
What GoZen adds
- Coordination: Reduce conflicts when multiple agents touch the same project.
- Observability: Track sessions, costs, errors, and activity in one place.
- Guardrails: Apply limits around spend, request rate, and sensitive actions.
- Task routing: Send different work to different providers or profiles.
Example configuration
{
"agent": {
"enabled": true,
"coordinator": {
"enabled": true,
"lock_timeout_sec": 300,
"inject_warnings": true
},
"observatory": {
"enabled": true,
"stuck_threshold": 5,
"idle_timeout_min": 30
},
"guardrails": {
"enabled": true,
"session_spending_cap": 5.0,
"request_rate_limit": 30
}
}
}
Common workflows
Multi-agent coordination
When several agents work in the same repository, GoZen can track file activity, surface warnings, and make collisions easier to avoid.
Session monitoring
Use the dashboard and APIs to inspect active sessions, token usage, error counts, and runtime duration.
Safety enforcement
Guardrails can pause runaway sessions, flag risky operations, and slow down retry loops before they become expensive.
Related docs
- Agent Infrastructure covers the newer runtime, observatory, coordinator, and guardrail architecture in more detail.
- Bot Gateway explains how to control running sessions from Telegram, Slack, Discord, and other chat platforms.
- Usage Tracking and Health Monitoring cover the metrics that power agent operations.