Skip to content
Agentic Levels
Start hereLevelsTracksPromptsAbout
Get my levelGet my level

  • Start hereYour first useful task with AI.
  • Level quiz8 questions. No AI needed.
  • Fluency ScoreYour Level and a score, from your own AI.

  • LevelsThe 11 levels, in order.
  • LessonsShort lessons, filter by level.
  • TracksGuided paths by tool, role or goal.
  • Set up your AI tool
  • Starter packFree files that help AI remember.

  • PromptsReady prompts to copy and adapt.
  • ChatGPT vs Claude vs GeminiPick the right one for the job.
  • How-tos
  • ReferenceCommands for Claude Code, Codex, Antigravity.
  • ResourcesHand-picked reading and tools.

  • Your progress
  • Your next moveAn interview that picks your next skill.
  • Your next lessonsThe next lessons at your level.
  • Updates

  • Pricing
  • About
  • FAQ & Feedback
  • Work with us
  • How we handle your data
Sign inPreferences
AboutPricingHow-tosPromptsStarter packMy stackHow we handle your data

© 2026 Fuentes Studio

PrivacyTermsRefundsAccessibilityAI content policyCredits
yourCouncil
Ready to help
✦

What do you want to understand?

Ask anything about what you're learning.

←How-to guides
  1. Resources
  2. ›How-to guides

How to use 12-Factor Agents

The 12 architectural principles behind reliable agent software.

Levels 3 to 7On GitHub

Your agent works in demos but breaks in production, unclear why. You need a mental model for what makes agents production-grade.

Before you start

Basic Claude Code sessions running

  1. 1Read the 12 factors
  2. 2Apply one factor at a time

Steps

01

Read the 12 factors

Clone or browse the repo. Each factor is a short, concrete principle. Read all 12 before doing anything else. They're designed to be read together, not picked from.

After reading: map your agent to the factors
I've read 12-Factor Agents. Walk me through each factor and tell me which ones my current setup violates. My setup: [AGENT_SETUP]. Give me a ranked list of which factors I'm furthest from and what to fix first.
[AGENT_SETUP]Your agent setup e.g. I use CLAUDE.md + A-MEM, run tasks in a ralph loop, no formal handoff protocol
02

Apply one factor at a time

Pick the factor you violate most. Implement it, verify it, then move to the next. Don't try to fix all 12 at once. The repo is designed for incremental adoption.

How do you know it worked?

  • You can name all 12 factors from memory
  • You've identified which 2-3 factors your current setup violates
  • You've implemented at least one factor and verified the improvement

If something looks off

Don't cherry-pick
The hard factors (idempotency, backpressure) are usually the most important, don't skip them for the easy wins.
Applies everywhere
The factors apply to any agent harness, not just Claude Code, use them wherever you build.

See also

  • Claude Code how-to
  • GSD: spec-driven workflow

What you'll do, 2 steps

  1. 1Read the 12 factors
  2. 2Apply one factor at a time

Was this helpful?