skills.rest
devopstales
devopstalesCOMMUNITY
Explore all skills by this author →
Website𝕏@_devopstalesGitHub

agentic-engineering

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

Updated Aug 27, 2026
Search Tags:#ai#agent#decomposition#engineering#model-routing#eval-first#cost-discipline
One-click install
npx skills add https://github.com/devopstales/ai-workflow --skill agentic-engineering-devopstales
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agentic-engineering
Source: https://github.com/devopstales/ai-workflow/tree/main/skills/agentic-engineering
Command: npx skills add https://github.com/devopstales/ai-workflow --skill agentic-engineering-devopstales

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing.

Core Features & Use Cases

  • Eval-first loop
  • Task decomposition into agent-sized units
  • Model routing by task complexity
  • Session management and milestone-based reviews
  • Review focus for AI-generated code: invariants, edge cases, security and risk
  • Cost discipline: track model, token estimates, retries, wall-clock time, success/failure

Quick Start

Decompose a task into agent-sized units and route steps by complexity while evaluating outcomes before execution.

Frequently Asked Questions about agentic-engineering

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

FAQPage Schema
How do I manage AI agent costs when routing model requests by task complexity?▼

Agentic engineering decomposes work into agent-sized, independently verifiable units to validate outcomes before committing changes. This ensures each step is evaluated against invariants and error boundaries during execution.

What is eval-first execution in AI agent workflows?▼

Eval-first execution validates outcomes before committing changes by running an evaluation loop. It enforces guardrails around security assumptions, invariants, and risk-aware reviews for AI-generated code prior to final execution.

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

Decompose large tasks into agent-sized, independently verifiable units to enable milestone-based reviews. This approach allows agentic workflows to validate individual outcomes before committing broader changes to the codebase.

Can I track token estimates and retries for agentic engineering sessions?▼

Track token estimates, retries, wall-clock time, and success/failure metrics for agentic engineering sessions. This cost discipline capability monitors per-task execution data alongside session management and milestone-based reviews.

When should I use cost-aware model routing instead of a single AI model?▼

Use cost-aware model routing when tasks have varying complexity levels requiring different model tiers. It enforces guardrails around invariants and error boundaries while validating outcomes before committing changes, unlike single-model approaches.

Related Skills

Explore More Software Engineering Skills→
yu13130122297yu13130122297

scout

Identify uncertainties in research framing, datasets, metrics, and baselines.

Community
Intermediate
ayourtchayourtch

subagent-driven-development

Dispatch fresh subagents per task with mandatory code reviews.

Community
Basic
nikhilvallishayeenikhilvallishayee
83

Checker

Detect edge cases and challenge assumptions through three verification questions.

Community
Basic
jfrometajfrometa

plan-writing

Decompose complex software work into small, verifiable tasks with dependencies.

Community
Basic