Skip to content
Get packs

Agentic Engineering Workflow

Guides an AI agent through eval-first development, task splitting, and model tier selection

Source Download skill
Repository
affaan-m/ECC
Stars
295,127
Licence
MIT
Files
1 file

What it does

This skill gives an AI coding agent a set of working habits for software projects where the agent writes most of the code and a human sets quality and risk rules. Before coding, the agent is told to define what 'done' means and write tests that capture the desired behavior, then compare results before and after. It also tells the agent to break large tasks into roughly 15-minute pieces, pick a cheaper or stronger model depending on difficulty, and track cost and retries. It is a process guide written as instructions; it does not add new tools or automate anything by itself.

Great for

  • Developers who use an AI agent for most implementation work and want more structured, testable output
  • Teams that want to track model cost and retries per task
  • Engineers reviewing AI-generated code who want a checklist focused on edge cases, error handling, and security assumptions

Skip it if

  • People who want a finished app or product without reviewing the code, since the skill assumes a human enforces quality
  • Projects where no tests or automated checks exist, because the eval-first loop depends on them
  • Beginners learning to code, since the skill presumes engineering experience and does not teach it

Try asking

  • “Plan how to break 'add user authentication' into small, independently testable units for my agent”
  • “Set up an eval-first loop for a new feature: write the failing test first, then implement and re-run”
  • “Review this AI-generated login module for edge cases, error handling, and security assumptions”
  • “Suggest which model tier (Haiku, Sonnet, or Opus) fits each step of this refactor and track cost per task”
Step-by-step guide: how to use Agentic Engineering WorkflowInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    agentic-engineering.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip agentic-engineering.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip agentic-engineering.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Provide your existing test suite or write evals before asking the agent to implement, so the baseline comparison is meaningful
  • Use the 15-minute unit rule as a starting point, then adjust to your project's size and risk areas
  • Confirm the model tier names available in your tool, since the skill refers to Haiku, Sonnet, and Opus by name

What's inside

  • SKILL.md

Redistributed unmodified from affaan-m/ECC under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.