TechPrufer
Get started
Cole Medin
Cole Medin logoCole Medin · Feature trail

Agentic Workflow Hooks

2 videos · Aug 2026

What this trail covers

This post continues the discussion on agentic workflows by highlighting the critical role of 'hooks' in ensuring reliability and adherence to rules within coding agents. Hooks are presented as guarantees for agent workflows, contrasting with mere guidance. The post details how to build and implement hooks, providing practical use cases and a 'hooks-create' skill for their generation. It emphasizes that hooks can enforce specific events and actions, thereby enhancing agent behavior and security. The discussion also touches upon the 'slop apocalypse' in AI-generated code, where comments and refactoring fall out of sync, leading to increased costs and reduced efficiency. The founder of BMAD advocates for a 'human on the loop' approach over full automation and discusses the evolution of AI coding frameworks, emphasizing planning and the development of self-improving AI skills. The importance of hiring systems thinkers and the need for multi-model strategies and effective prompting are also explored.

All Cole Medin trails
  1. Watch This If Your Coding Agent is Ignoring Your Rules (You Need Hooks)This post introduces and advocates for the use of 'hooks' in coding agents. It explains that hooks act as guarantees for agent workflows, ensuring that certain actions or events are always handled, unlike standard rules which can be ignored. The author demonstrates how to build hooks, provides a 'hooks-create' skill to automate their generation, and offers practical examples for their application, including guaranteeing test execution, enhancing security, and ensuring audit trails. The core contribution is the elevation of hooks from an overlooked feature to a fundamental component for reliable agentic systems.
  2. BMAD's Founder on the Future of AI Coding (And the Slop Apocalypse)This post contributes to the understanding of agentic workflows by introducing the concept of 'hooks' as a mechanism for enforcing reliability and rules in coding agents. It details the implementation and use cases of these hooks, including a 'hooks-create' skill. The post also addresses the long-term maintenance challenges of AI-generated code, termed the 'slop apocalypse,' and advocates for a 'human on the loop' approach. It discusses the evolution of AI coding frameworks, the shift towards planning, the development of self-improving AI skills, and the importance of systems thinking and multi-model strategies in AI development.