What problem does it solve?
This skill addresses the lack of structure in AI-assisted development by enforcing a disciplined, eval-first approach that minimizes wasted effort and reduces implementation risk.
Core Features & Use Cases
- Eval-First Execution: Ensures every code change is validated against baseline and regression tests before and after implementation.
- Task Decomposition: Breaks complex engineering requirements into verifiable 15-minute units to maintain focus and clarity.
- Model Routing: Optimizes cost and performance by matching specific engineering tasks to the appropriate model tier (Haiku, Sonnet, or Opus).
- Use Case: When refactoring a legacy module, use this skill to define success criteria, decompose the refactor into testable units, and route the architectural analysis to the most capable model.
Quick Start
Activate the agentic-engineering skill to begin a new development session by defining the completion criteria and regression tests for your current task.