Twisted Brackets Blog

Production Agentic AI: Permissions, Observability, and Kill Switches

Putting an agentic AI system into production is a different engineering problem to putting it into staging. Once it is live, the governance question shifts from "will it behave?" to "how do we know if it stops behaving, and what do we do about it?" A practical guide to production permissions, observability architecture, and kill switches for Australian engineering leaders in regulated industries.

25 Sept 2026

17 min read

What Australian CTOs Should Put in the AI Governance Pack

// AI;21 Sept 202616 min read

Governing Agentic AI: Human-in-the-Loop Patterns When Tools Can Act (MCP Edition)

// AI;16 Sept 202614 min read

Vercel Can Do WebSockets Now. Here's Where a Chat Room Still Breaks It.

// Next.js;15 Aug 202611 min read

AI-First Programming Languages: What They Are, and Whether You Should Use One Yet

// AI;9 Aug 202614 min read

lucide-animated: Icon Motion That Knows When to Stop

// JavaScript;2 Aug 20268 min read

Animated Gradient Borders: A Better Loading State for AI Interfaces

// CSS;31 July 202612 min read

Pathfinding Algorithms Explained

// Game development;26 July 20268 min read

Tokens, Vectors, and Attention: How an LLM Actually Reads Your Prompt

// AI;26 July 202612 min read

SOLID Principles Matter More With AI Coding Agents, Not Less

// AI;25 July 202611 min read