How I Build Agentic Systems

Core Tooling

  • Hermes skills and coding agents for orchestration
  • n8n for workflow automation and integrations
  • Local and hosted LLMs for routing, summarisation, and execution
  • NetBox, Ansible, and scheduled jobs for infrastructure control
  • Prometheus, Grafana, and centralised logs for observability

Guardrails

  • Dry-run first for risky actions
  • Tiered approval for destructive changes
  • Clear audit trails and timestamped outputs
  • Verification after change, not just execution
  • Monitoring for both agent health and business impact

Design Principles

I build agentic systems as operational software, not as demos. That means narrow responsibilities, observable outputs, explicit rollback paths, and enough structure that the system is safe to run repeatedly in production.