agentic-engineering

Coordinate AI agents to implement engineering tasks with eval-first validation.

2|2|Updated Mar 12, 2026

Search Tags:#engineering #agentic #task-decomposition #model-routing #eval-first #cost-discipline

Ready to run

npx skills add https://github.com/zhmxiaowo/opencode-simple --skill agentic-engineering-zhmxiaowo

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agentic-engineering
Source: https://github.com/zhmxiaowo/opencode-simple/tree/main/.opencode/skills/agentic-engineering
Command: npx skills add https://github.com/zhmxiaowo/opencode-simple --skill agentic-engineering-zhmxiaowo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps teams coordinate AI agents to perform engineering work with eval-first checks, decomposition, and cost-aware routing to improve reliability and governance.

Core Features & Use Cases

Eval-first loop to baseline capability and regression checks.
Task decomposition using the 15-minute unit rule for independently verifiable units with a single dominant risk.
Model routing to assign work to Haiku, Sonnet, and Opus tiers based on task complexity.
Cost discipline with per-task tracking of model, tokens, retries, and wall-clock time.

Quick Start

Plan and execute a small engineering task by decomposing it into verifiable units and routing work by model tier to maximize quality and speed.

Frequently Asked Questions about agentic-engineering

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

FAQPage Schema

What is an eval-first loop for coordinating AI agents in software engineering?▼

An eval-first loop baselines AI agent capability and runs regression checks before implementation. This approach ensures engineering tasks yield verifiable results through unit-level checks, preventing regressions when scaling from small features to multi-file systems.

How do I decompose engineering tasks for AI agents into verifiable units?▼

Decompose engineering tasks using the 15-minute unit rule to create independently verifiable units. Each unit should have a single dominant risk, allowing AI agents to implement work reliably and ensuring outcomes are checked through unit-level validation.

How does model routing work for AI agents implementing engineering tasks?▼

Model routing assigns engineering work to Haiku, Sonnet, and Opus tiers based on task complexity. This cost-aware strategy maximizes quality and speed by matching the AI agent's model tier to the specific requirements of the decomposed unit.

How do I track token usage and costs when coordinating AI agents for coding?▼

Track costs by monitoring per-task metrics including model tier, token estimates, retries, and wall-clock time. This cost discipline ensures governance and transparency when scaling AI agents across multi-file engineering systems.

Can I use this agentic engineering approach for both small features and multi-file systems?▼

Yes, this approach scales from small features to multi-file systems by applying risk-based review checkpoints. It ensures verifiable results through regression evaluation while maintaining cost discipline across different levels of engineering complexity.

What are the limitations of using cost-aware model routing for AI engineering tasks?▼

Cost-aware model routing requires balancing quality and speed across Haiku, Sonnet, and Opus tiers. Limitations include managing risk-based review checkpoints and ensuring regression evaluation catches errors when routing complex multi-file engineering tasks.