The Challenge of Securing AI Agents: A DevSecOps Perspective

As a DevSecOps Team Leader, my job is to secure customers using modern technologies. Sounds straightforward, right? The reality is far more complex. Every day, I face the challenge of enabling innovation while maintaining security. The rapid adoption of AI has introduced a new dimension to this challenge: agentic AI assistants that do not just chat, they act. This challenge connects directly to something I wrote about recently. In my article on spec-driven development with GitHub Spec-Kit, I discussed how structure and governance matter when using AI for coding. The same principle applies here: when AI agents can execute code, access secrets, and operate with user privileges, we need structure and governance more than ever. ...

February 17, 2026 · 5 min · 1059 words · Matteo Bisi

Testing Spec-Kit: Building a Functional Container TUI in 2.5 Hours

Introduction: Theory Meets Practice In my previous article about GitHub Spec-Kit, I explored the theoretical foundations of spec-driven development: why structured AI workflows matter for compliance, auditability, and team collaboration. I discussed the high-level concepts of audit trails, liability, and how spec-kit transforms “vibe coding” into a rigorous, documented process. Today, I’m sharing something different: a raw, unfiltered hands-on experience building a real tool from scratch using spec-kit. This is a chronological journey documenting what actually happened when I let spec-kit drive the development process from constitution to working code. ...

February 12, 2026 · 9 min · 1747 words · Matteo Bisi

AI CLI Standardization: From Tool Lock-in to Portability

Introduction: From Web Chatbots to CLI Tools AI is a powerful tool, and for IT professionals, the most effective way to leverage it is through CLI tools like GitHub Copilot CLI, Claude Code, Gemini CLI, or similar agents. In previous articles like GitHub Spec-Kit, I explored spec-driven development and structured AI workflows, but I realized I skipped fundamental concepts: why CLI tools beat web chatbots and how to standardize your AI setup for portability. ...

February 6, 2026 · 12 min · 2506 words · Matteo Bisi

When Your Update System Becomes the Attack Vector: The Notepad++ Supply Chain Compromise

The recent Notepad++ supply chain compromise shows how even widely trusted, open-source tools become vectors for state-sponsored espionage when their distribution infrastructure falls into the wrong hands. This was a surgical, six-month operation that bypassed traditional code security controls by exploiting the update mechanism. What Happened and Where the SDLC Failed In 2025, Notepad++, a widely used open-source text editor, suffered a sophisticated supply chain attack. Chinese state-sponsored threat actors compromised the shared hosting provider in June, gaining control of the update distribution system. Even after losing direct server access in September following a kernel update, attackers maintained persistence through stolen credentials until December 2. The fixed version 8.8.9 with hardened update verification was released on December 9. ...

February 3, 2026 · 7 min · 1370 words · Matteo Bisi

ClawdBot → MoltBot → OpenClaw: A Case Study in Confusion Attacks and Security Risks

What is ClawdBot/MoltBot/OpenClaw? For those unfamiliar with the project, OpenClaw (formerly MoltBot, previously ClawdBot) is a personal AI assistant platform that integrates with multiple messaging channels including WhatsApp, Telegram, Discord, Slack, Signal, iMessage, and many others. The project is available at github.com/openclaw/openclaw and maintains a website at openclaw.ai. The tool is designed to be a “local-first, single-user assistant” with capabilities that include shell command execution, filesystem operations, browser automation, and integration with various cloud services. It’s essentially a bridge between AI models and your entire digital ecosystem. However, OpenClaw does not provide model access itself; users must configure it with their own API keys from providers like Anthropic, OpenAI, or others. ...

January 31, 2026 · 11 min · 2145 words · Matteo Bisi

GitHub Spec-Kit: Why Structured AI Development Beats Vibe Coding

Introduction: Spec-Driven Development vs. Vibe Coding If you’ve been working with AI coding assistants, you’ve probably experienced what some call “vibe coding”, throwing prompts at an LLM and hoping for the best. Sometimes it works brilliantly. Other times, you end up with code that technically runs but doesn’t align with what you actually needed, or worse, introduces architectural decisions that create technical debt down the road. Spec-Driven Development (SDD) flips this approach on its head. Instead of starting with code and documenting later (if at all), you begin with comprehensive specifications that define the what and why before anyone, human or AI, writes a single line of code. The specification becomes the single source of truth, guiding implementation and ensuring alignment across the entire team. ...

January 21, 2026 · 6 min · 1267 words · Matteo Bisi

Evaluating Oss Security Fresh Editor s2c2f

It’s December 27th, and like most of you, I’m somewhere between “fully checked out for the holidays” and “can’t stop tinkering with new tools on my laptop.” Nobody’s at work. Teams is shut down and Slack is quiet. The corporate VPN can wait until January. But my curiosity? That’s working overtime. A couple of weeks ago, I discovered Fresh, a Rust-based terminal text editor that feels like it was designed specifically for people like me who live in terminals. Here’s what caught my attention: ...

December 27, 2025 · 10 min · 2034 words · Matteo Bisi

2025 CWE Top 25: Mitre's Critical Software Weakness Rankings and Trends

MITRE released the 2025 CWE Top 25 on December 11, 2025, identifying the most dangerous software weaknesses based on 39,080 CVE Records published between June 2024 and June 2025. The list ranks weaknesses by their frequency as root causes in CVE data and their CVSS severity scores, highlighting persistent threats like XSS and SQL Injection alongside emerging issues such as authorization flaws and memory bugs—key priorities for DevSecOps teams securing modern cloud‑native applications. Explore how the 2025 rankings differ from 2024, the top ten shifts, and what CWE root causes reveal beyond CVE trends. ...

December 17, 2025 · 6 min · 1104 words · Matteo Bisi

Kubernetes Security: 2025 Stable Features & 2026 preview

Like your favorite music streaming service’s 2025 Wrapped®, here’s my recap of Kubernetes security highlights from 2025, plus predictions for features likely graduating to stable in early 2026. As a DevSecOps Team Leader, I bridge development speed with security rigor daily. Kubernetes and cloud-native security are my passion, especially hardening workloads for production. With Kubernetes v1.35 releasing December 17, now’s the perfect time to review 2025’s security wins and plan for 2026. ...

December 8, 2025 · 4 min · 707 words · Matteo Bisi

Back to Basics: My Opinionated 2025 sshd_config Hardening

In today’s fast-paced tech landscape, it’s common to find incredibly talented engineers mastering complex orchestrators like Kubernetes or crafting intricate Infrastructure as Code solutions. We’re living in an era of high-level abstractions, which is fantastic for productivity. However, this focus on the ’new and shiny’ can sometimes lead us to overlook the foundational bedrock upon which everything is built. It might seem a bit old-school to write a blog post about hardening SSH in 2025. Yet, these ‘basic’ skills are more critical than ever. In a world of ephemeral infrastructure and complex supply chains, securing the front door to our systems remains a non-negotiable first step. ...

December 3, 2025 · 8 min · 1625 words · Matteo Bisi