TL;DR: The last two posts wired detection to NetClaw and then made it cheap to run. This one goes the other direction: instead of the network waking NetClaw up, I edit a device interface in Nautobo...
TL;DR: In the last post I wired firewall alerts to NetClaw. This post is what happens on the other side of that webhook, and how I stopped it from being expensive. NetClaw injects a one-line summar...
TL;DR: My pfSense firewall was great at blocking traffic and terrible at telling me when something slipped through. So I built detection. Not with a new bespoke agent like I tried back in Part 5, b...
▶️ Watch the video Watching a network engineer’s eyes light up in real time watching an AI agent build a full service provider core from nothing. No slides. No pre-baked configs. Just a blank prom...
In Part 14, I made NetClaw a BGP peer. I injected a prefix, withdrew it, watched the control plane react. Great theater. Wrong monitoring plane. The Grafana dashboard I’d built for “route stabilit...
▶️ Watch the video In Part 13, we deployed the observability stack — OTEL Collector polling SNMP, VictoriaMetrics storing metrics, Loki aggregating syslog, Grafana rendering dashboards. NetClaw ca...
▶️ Watch the video In Parts 1 through 6, I built an observability pipeline on my home network: OpenTelemetry Collector polling SNMP, VictoriaMetrics storing time-series, Loki aggregating syslog,...
▶️ Watch the video I have a demo. One prompt, one conversation, and the agent builds a complete network lab — 20 routers and switches, source of truth, generated configurations, compliance monitor...
▶️ Watch the video Nautobot is a source of truth. It knows what devices you have, what interfaces they have, what VLANs are assigned, what IP addresses are allocated. But it doesn’t know what your...
I didn’t know what spec-driven development was three weeks ago. I’d never written a spec.md before writing code. I’d never asked an API “what fields do you actually have?” before designing an integ...