Agency Law Your Tesla Changed While You Slept - So Did Your Duty of Reasonable Care An over-the-air update changed your car's braking overnight. It also changed your legal duties. No doctrine has a name for that relation — agency, products liability, bailment all fail — and it now sits under every LLM agent deployment. By John F. Holliday • 10 min read
Agentic AI Your Chatbot Is Not a Separate Legal Entity Air Canada called its chatbot a separate legal entity. Agency law disposed of that argument eight centuries ago — and it governs your AI agents today. By John F. Holliday • 6 min read
Agentic AI New Book: The Law of Agents Eight centuries of doctrine vs. eighteen months of AI. By John F. Holliday • 4 min read
Agentic AI Two Algorithms, Zero Shared Memory When prior authorization meets retrospective denial, the real failure isn't procedural — it's semantic. By John F. Holliday • 11 min read
Agentic AI Semantic Dissonance: The Silent Failure Mode of Multi-Agent AI Systems When agents share a channel but not a contract, coherence collapses without warning — and domain-specific languages may be the only reliable remedy. By John F. Holliday • 11 min read
Agentic AI When Your AI Becomes Chief of Staff: A STRIDE Threat Analysis of the Lobster/OpenClaw Personal AI Agent System I constructed a complete data flow diagram with five trust boundaries and ran a systematic STRIDE-per-element analysis. The result: 46 enumerated threats — 11 Critical, 17 High, 14 Medium, and 4 Low. Seven have no mitigation at all. By John F. Holliday • 8 min read
Agentic AI When Your AI Agent Becomes the Attack Surface: The OpenClaw Security Crisis and What It Means for All of Us The fastest-growing open-source project in GitHub history is also 2026's most significant cybersecurity incident. Here's what happened, why it matters, and how to protect yourself. By John F. Holliday • 13 min read
Agentic AI Most AI Agents Are Just Fancy Prompt Wrappers. I Built One That Actually Understands Its Own Output Grammar-validated AI generation with language server infrastructure: AI systems that reason about structured domains with the same rigor as a compiler. By John F. Holliday • 5 min read
Two Robots Examining a Document for Compliance Agentic AI Instinct vs. Deliberation: How Anthropic and OpenAI Train Their Models to Follow the Rules — And Why It Matters for Enterprise AI The most consequential technical distinction in enterprise AI isn't about which model is smarter. It's about how each model was taught to be safe — and what that means when you deploy agents that make real-world decisions. By John F. Holliday • 12 min read
MCP Contract Enforcement Agentic AI MCP Needs a Type System, Part 2: Building the Contract Layer Tool descriptions suggest constraints to LLMs, but suggestions aren't guarantees. So what would formal MCP contracts actually look like? By John F. Holliday • 7 min read
AI Agents Agentic AI MCP Needs a Type System, Part 1: Six Incidents That Expose the Protocol's Blind Spot Your AI agents are only as safe as the contracts they honor—and right now, MCP doesn't have any. By John F. Holliday • 6 min read
Agentic AI Deontic Logic for Agent Permissions: A Formal Framework for AI Agent Governance The fundamental question of what agents are permitted to do remains governed by ad-hoc JSON schemas and vibes-based access control. Wesley Hohfeld's decomposition of rights into eight fundamental relations—and deontic logic's formal operators—provide exactly the rigor AI agent governance needs. By John F. Holliday • 8 min read
