Free · Guide
A 30-day AI learning plan for technical literacy.
An hour a day. One practice repo. Let the agent build; you direct it and review each PR.
Week 1
Days 1–4
Version control
Create the repo, turn on branch protection, ship small changes through PRs, resolve a conflict you plant, run two agents in worktrees.
Done whenFive PRs merged through required checks.
Days 5–9
Web and APIs
Write an OpenAPI contract first, generate a mock server, build two endpoints, call them with curl, add a signed webhook.
Done whenYou can explain any request in the Network tab.
Week 2
Days 10–13
Data
Turn a domain model into a schema, write three migrations, add an index, seed 100,000 rows, restore a backup.
Done whenA restore to a fresh database works.
Days 14–18
Identity and security
Add an auth provider, two roles and two tenants with RLS; run /security-review; plant a prompt injection in a test document.
Done whenThe cross-tenant test fails safely and the injection is ignored.
Week 3
Days 19–22
Infrastructure
Move config to environment variables and secrets to a manager, add a dev container, deploy preview and staging on HTTPS, add one queued job.
Done whenA fresh clone runs in one command; staging is live.
Week 4
Days 23–26
Quality
CI with types, lint, tests and accessibility checks; 20 eval cases for the agent feature; one load test.
Done whenCI blocks a deliberately broken PR.
Days 27–28
Operations
Structured logs, one trace, one alert and one feature flag; rehearse a rollback.
Done whenFeature off and rollback done in under five minutes.
Days 29–30
Cost and review
Read /usage, work out cost per task card, write a one-page census of the app.
Done whenYou can explain its architecture, risks and monthly cost to a colleague.
After day 30
Each week, pick one concept from a real PR and ask the agent to teach it to you on your own code.
Source: The Agentic Studio Playbook, Chapter 41, Table 41.9