What is the focus of the AI-generated code review?

The review focuses on high-risk areas like security assumptions, error boundaries, and hidden architectural coupling rather than stylistic choices that are handled by automated formatters.

What is the core principle of Agentic Engineering?

The core principle is an evaluation-first approach where success criteria and regression tests are established before the AI agent begins implementation to ensure quality and prevent regressions.

Which models does this skill recommend for specific tasks?

It routes tasks based on complexity: Haiku for classification and boilerplate, Sonnet for implementation and refactoring, and Opus for high-level architecture and root cause analysis.

How does the 15-minute unit rule work?

It requires breaking down complex tasks into smaller units that can be completed and verified within 15 minutes, ensuring each unit has clear completion conditions and localized risk.

How does this skill help manage AI development costs?

It implements cost discipline by tracking model usage and tokens, only escalating to higher-tier models (like Opus) when lower-tier models (like Sonnet or Haiku) fail to solve a reasoning gap.

Agentic Engineering

Name: Agentic Engineering
Author: affaan-m

byaffaan-m

•

137,500

•

Productividad y Flujo de Trabajo

Optimizes AI agent workflows through evaluation-first cycles, task decomposition, and cost-aware model routing.

Agentic Engineering provides a structured framework for managing AI implementation where humans transition into quality and risk control roles. By implementing a rigorous 'evaluation-first' cycle, the skill ensures that completion criteria and regression tests are defined before any code is written. It leverages the '15-minute unit rule' for granular task decomposition and features an intelligent routing system that assigns tasks to Haiku, Sonnet, or Opus based on complexity. This skill is essential for developers seeking to scale their AI-driven development while maintaining strict control over code quality, security, and operational costs.

Características Principales

01Evaluation-first cycles that define success criteria and regression checks before implementation.

02Granular task decomposition using the 15-minute unit rule for manageable agent workflows.

03Security-centric code reviews focusing on edge cases, boundaries, and hidden coupling.

04Strategic session management to maintain context while preventing token bloat.

05Cost-aware model routing across Haiku, Sonnet, and Opus to optimize performance and spend.

06137,500 GitHub stars

Casos de Uso

01Implementing new features with built-in capability assessments and baseline regression testing.

02Optimizing development costs by automating boilerplate tasks with lower-tier models.

03Executing complex architectural refactors across multiple files using tiered model routing.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code agentic-engineering

For use in Claude.ai and ChatGPT

Características Principales

01Evaluation-first cycles that define success criteria and regression checks before implementation.

02Granular task decomposition using the 15-minute unit rule for manageable agent workflows.

03Security-centric code reviews focusing on edge cases, boundaries, and hidden coupling.

04Strategic session management to maintain context while preventing token bloat.

05Cost-aware model routing across Haiku, Sonnet, and Opus to optimize performance and spend.

06137,500 GitHub stars

Casos de Uso

01Implementing new features with built-in capability assessments and baseline regression testing.

02Optimizing development costs by automating boilerplate tasks with lower-tier models.

03Executing complex architectural refactors across multiple files using tiered model routing.