What is the primary goal of the Agentic Engineering skill?

It aims to provide a structured framework where AI agents handle the bulk of implementation while humans focus on high-level quality and risk control.

Does this skill help manage development costs?

Yes, it includes cost discipline protocols that track model usage, token estimates, and retries, only upgrading to higher-tier models when lower tiers fail.

How does the 15-minute unit rule work?

It breaks tasks into small, independently verifiable units that address one primary risk and have a clear, exposure-based completion condition.

How does this skill handle model routing?

It assigns tasks to Haiku for classification and boilerplate, Sonnet for implementation and refactoring, and Opus for architecture and root cause analysis.

What is the 'Evaluation-First' cycle?

It is a process where capabilities and regression assessments are defined and baselined before any code implementation begins to ensure measurable results.

Agentic Engineering

Name: Agentic Engineering
Author: oabdelmaksoud

byoabdelmaksoud

0•

生产力与工作流

Manages AI-driven development workflows through evaluation-first execution, task decomposition, and cost-optimized model routing.

The Agentic Engineering skill transforms Claude into a disciplined engineering partner by implementing a structured workflow for AI-assisted development. It prioritizes quality control through pre-execution evaluation, breaks complex tasks into manageable 15-minute units, and optimizes costs by routing tasks to the most appropriate AI model layer—Haiku for simple edits, Sonnet for implementation, and Opus for complex architecture. This skill is ideal for engineering teams looking to scale implementation using AI agents while maintaining rigorous standards for code reviews, error handling, and operational efficiency.

主要功能

010 GitHub stars

02Cost-Aware Model Routing: Automatically selects the most efficient model (Haiku, Sonnet, or Opus) based on complexity.

03Risk-Focused Code Review: Directs human oversight toward boundary cases, security, and architectural coupling.

04Evaluation-First Cycle: Defines success criteria and regression tests before code execution.

05Session Strategy Management: Optimizes context windows through milestone compression and session transitions.

06Intelligent Task Decomposition: Breaks projects into 15-minute units with clear completion conditions.

使用场景

01Optimizing API costs and token usage in large-scale AI engineering projects.

02Scaling software development capacity using autonomous AI agent workflows.

03Implementing rigorous quality control for AI-generated code in production environments.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add oabdelmaksoud/agention agentic-engineering

For use in Claude.ai and ChatGPT

主要功能

010 GitHub stars

02Cost-Aware Model Routing: Automatically selects the most efficient model (Haiku, Sonnet, or Opus) based on complexity.

03Risk-Focused Code Review: Directs human oversight toward boundary cases, security, and architectural coupling.

04Evaluation-First Cycle: Defines success criteria and regression tests before code execution.

05Session Strategy Management: Optimizes context windows through milestone compression and session transitions.

06Intelligent Task Decomposition: Breaks projects into 15-minute units with clear completion conditions.

使用场景

01Optimizing API costs and token usage in large-scale AI engineering projects.

02Scaling software development capacity using autonomous AI agent workflows.

03Implementing rigorous quality control for AI-generated code in production environments.