Technical Writing
& Reflections
Stories behind the systems. Errors that became insight. And commands that stuck with me.
🔄 GitOps for Docker Compose: Doco-CD, SOPS, and Renovate
October 29, 2026
What happens when the deployer passes its healthcheck but stops deploying? Standardizing Docker Compose across eight targets and 15 application stacks with Doco-CD, SOPS, Renovate, and Gitea Actions: per-target canary updates, scoped secrets, and recovery that fixes Git before repairing the agent.
☁️ tfSTRIDE: Evolving Terraform Threat Modeling into a Multi-Cloud Engine
June 22, 2026
tfSTRIDE has grown from AWS-first plan review into a provider-aware analysis system spanning deeper AWS coverage, near-parity GCP support, and scoped AzureRM fundamentals. This case study covers the plugin architecture, normalized inventory model, trust-boundary engine, report contracts, dashboard reuse, and CI gating workflow that made that expansion possible.
🔎 tfSTRIDE: Rule-Based Terraform Threat Modeling Before Deployment
May 2, 2026
Terraform plans are readable, but still easy to misjudge when IAM trust, network posture, and private data access interact. I built tfSTRIDE to turn plan JSON into trust-boundary analysis, explainable findings, and CI-ready policy-gating output before deployment.
🛰️ Drift‑Detection And Fleet‑Wide Enforcement
March 10, 2026
Over time, I noticed significant drift accumulating between CrowdSec LAPI decisions and nftables enforcement across my fleet. To solve it, I built a backend agent that autonomously detects and heals drift, and a frontend dashboard that visualizes enforcement state across all nodes in real time.
🧠 A Practical Mini-SIEM
January 22, 2026
Modern security stacks often generate endless logs, but very few clear decisions. This is a walkthrough of building a lightweight, distributed threat-decision engine focused on confidence, severity, and explainability. Heavyweight tooling isn't always required to make high-quality security decisions.
⚙️ Automated Node Health Failover
January 10, 2026
Automated DNS failover with minute-level health checks across five production endpoints. Built with Python, Cloudflare's API, and a <$7/year VPS - detecting failures in real-time and switching traffic between primary and mirror nodes without manual intervention.