“Agentic Exposure” Is Not the Same as Agent Engineering
Why experience with the latest agent framework matters less than understanding context, cost, permissions and deterministic execution boundaries.
Ai
When an Enterprise AI Agent Goes Rogue
A Rovo agent created a Jira issue in the wrong project—revealing why enterprise AI needs deterministic context boundaries, not probabilistic guesswork.
The "Agentic Exposure" Checklist: SDK Churn, Context Taxes, and Real Systems
Why treating bleeding-edge developer kits like Google ADK as enterprise prerequisites misses the point: rapid SDK churn, the compounding context feeding tax in tools like Claude Code, and why Atlassian Rovo proves we are still in the early days.
TDD for AI Agents: Why Atlassian Rovo Fails and How to Fix It
Why enterprise agents like Atlassian Rovo create tasks in the wrong project, the dangers of ambient write permissions, and how deterministic TDD test suites prevent autonomous failures.