agentic-engineering

Coordinate AI-driven engineering workflows with eval-first execution and cost-aware model routing.

16|3|Updated Mar 12, 2026

Search Tags:#decomposition #engineering #agent-based #model-routing #eval-first #cost-discipline

Ready to run

npx skills add https://github.com/sehoon787/my-claude --skill agentic-engineering-sehoon787

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agentic-engineering
Source: https://github.com/sehoon787/my-claude/tree/main/skills/ecc/agentic-engineering
Command: npx skills add https://github.com/sehoon787/my-claude --skill agentic-engineering-sehoon787

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Enable engineering workflows where AI agents perform most implementation work while humans enforce quality and risk controls.

Core Features & Use Cases

Eval-first loop: define capability eval and regression eval, run baseline, then execute implementation and re-evaluate deltas.
Task decomposition: break work into agent-sized units with clear completion criteria.
Model routing: assign Haiku for classification/boilerplate transforms, Sonnet for implementation/refactors, and Opus for architecture/root-cause analysis.
Cost discipline: track model tier, token usage, retries, and wall-clock time to optimize efficiency.

Quick Start

Begin by outlining a small engineering task, decomposing it into agent-sized steps, selecting the appropriate model tier, and running the eval-first loop.

Frequently Asked Questions about agentic-engineering

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I route AI tasks to different models based on complexity?▼

Route AI tasks by assigning Haiku for classification or boilerplate transforms, Sonnet for implementation or refactors, and Opus for architecture or root-cause analysis. This model routing strategy balances capability against cost for each agent-sized unit of work.

What is an eval-first loop for AI engineering workflows?▼

An eval-first loop defines capability and regression evals, runs a baseline, executes implementation, and re-evaluates deltas. This process ensures AI-driven engineering workflows maintain quality through iterative evaluation from design to implementation.

How do I decompose engineering tasks into agent-sized units?▼

Break engineering tasks into agent-sized units with clear completion criteria. Task decomposition allows AI agents to perform implementation work while humans enforce quality controls and monitor risk throughout the staged workflow.

How do I track token usage and wall-clock time for AI agent workflows?▼

Track token usage, retries, and wall-clock time across model tiers to maintain cost discipline. Monitoring these metrics alongside model routing decisions optimizes efficiency and exposes operational costs for AI-driven engineering workflows.

Can I use this approach for large-scale refactoring projects?▼

Yes, staged decomposition scales refactoring by assigning Sonnet for implementation and Opus for root-cause analysis. The eval-first loop regression tests ensure that large-scale changes maintain capability baselines throughout the iterative process.

Why use eval-first execution instead of direct AI implementation?▼

Eval-first execution prevents regression by establishing baselines before implementation. Direct AI implementation risks quality degradation, whereas iterative evaluation enforces human quality and risk controls over agent-generated code.