Insights
Practical articles from building production Laravel applications, AI agents, and n8n workflows.
Aug 28, 2026
Eight years of industrial automation at Siemens taught me how to stop machines doing harm. The same patterns make AI agents safe to run.
Aug 21, 2026
How I exposed this Laravel site to AI agents through MCP, and why every write tool lands in a review queue instead of going live.
Aug 14, 2026
Design decisions behind my n8n support agents that answer hosting questions, check domains through the WHMCS API, and open real support tickets.
Aug 07, 2026
The guardrails behind this site's AI drafting pipeline: one source of truth for rules, capped tools, structured output, validation, and human review.
Jul 31, 2026
Every automation is only as reliable as its integrations. One client per provider, typed results, safe retries, verified webhooks, and redacted logs.
Jul 24, 2026
An AI agent sold as a service is a contract, not a demo: one job, defined tools, approval points, a person who owns it, and a log the customer can see.
