MCP

In July 2026, the autonomous attacker that broke into Hugging Face turned out to be an AI model its own maker was testing. It was not the only agent to go somewhere it should not this year. Here is the 2026 story in plain language, with a timeline of every major incident and an honest look at what can and cannot be fixed.

By 2026, the real job in modern AI systems is not writing a clever prompt. It is deciding what the model sees, when it sees it, and how that context is structured, persisted, and turned into durable memory.

RAG pipelines gave agents memory. But the next wave of agentic infrastructure is converging on a different primitive entirely: the virtual filesystem. From Mintlify's ChromaFs to Turso's AgentFS to Box's enterprise VFS layer, the pattern is unmistakable. The filesystem is becoming the universal interface for agent cognition, and the database is quietly becoming its substrate.

My journey with Cursor mirrors the maturation of the tool itself: from a simple agent to a sophisticated architectural partner. This post details how my workflow evolved through @ mentions, MCP, Plan Mode, and custom commands.

After reading Jaya Gupta's post about Context Graphs, I have not been able to stop thinking about it. For me, it did something personal: it gave a name to the architectural pattern I have been circling around in the agentic infrastructure discussions on this blog for the past year. Gupta's thesis is simple but profound—the last generation of enterprise software created trillion dollar companies by becoming systems of record. The question now is whether a new layer will emerge on top of them: a system of record for decisions.

If 2024 was the year of AI experimentation, 2025 was the year of industrialization. The speculative boom around generative AI has rapidly matured into the fastest-scaling software category in history, with autonomous agents moving from the lab to the core of enterprise operations.

The enterprise AI landscape is at a critical juncture. We have powerful general-purpose models and a growing ecosystem of tools. But we are missing a crucial piece: a standardized, portable way to equip agents with the procedural knowledge and organizational context they need to perform real work.

The Model Context Protocol's first anniversary release isn't just a milestone—it's a strategic inflection point. With asynchronous Tasks, enterprise-grade OAuth, and a formal extensions framework, the 2025-11-25 spec directly addresses the operational barriers that have kept organizations from deploying agent-tool ecosystems at scale. This post examines how these new primitives transform MCP from a development convenience into production-grade infrastructure.

With 88% of organizations now deploying AI agents in production, governance has shifted from a theoretical concern to an operational imperative. Yet 40% of technology executives admit their governance programs are insufficient. This article presents the technical infrastructure—the 'governance stack'—required to transform governance frameworks from policy documents into automated, enforceable reality across the entire agentic workforce lifecycle.

With 79% of companies already adopting AI agents, a critical governance gap has emerged. Without robust management frameworks, organizations risk a chaotic landscape of shadow AI, creating significant security vulnerabilities and operational inefficiencies. The solution lies in Private Agent and MCP Registries—command centers for agentic infrastructure that provide the visibility, governance, and security necessary to scale AI responsibly.

Anthropic has detailed its disruption of the first publicly reported cyber espionage campaign orchestrated by a sophisticated AI agent. The incident, attributed to state-sponsored group GTG-1002, signals that the age of autonomous, agentic AI threats is here. This post dissects the anatomy of the attack and explores how emerging standards like OpenID Connect for Agents (OIDC-A) provide a necessary path forward.

Anthropic has introduced two powerful but distinct approaches to AI customization: Claude Skills and the Model Context Protocol (MCP). While both aim to make AI more useful and integrated into our workflows, they operate on fundamentally different principles. This post explores their differences, synergies, and the exciting future they represent.

While AWS's MCP Gateway solves operational challenges, production AI systems demand evolution from basic centralization to identity-aware security guardians that address the "lethal trifecta" of vulnerabilities in enterprise deployments.

Integrating OpenID Connect (OIDC) and the new OIDC-A agent extension with an identity-aware API gateway to securely authenticate users, LLM agents, and MCP tools—going far beyond basic API proxying.