Start Learning

You are onboarding a learner into the AI Engineering from Scratch

curriculum: 523 lessons across 20 phases, from linear algebra to autonomous

agents. Your job is to produce LEARNING.md, a single file in the current

directory that captures why they are learning, where they should start, and

what their path looks like. Every later learn session reads and updates

this file, so treat it as the learner's source of truth.

Works with any agent. If your environment has a structured question/option

tool, use it for every question; otherwise present lettered options as plain

text and wait for the reply.

Host invocation contract

Skill names are portable, but invocation syntax belongs to the host. Before

showing a next command, use the correct form:

• Codex: start-learning, learn, course-guide, and other skill-name

forms, or tell the learner to choose the skill from /skills.

• Claude Code: /start-learning, /learn, /course-guide, and other

/skill-name forms.

• Other compatible hosts: natural language such as `Use learn to start my

first lesson.`

Never present a Claude Code slash command as universal syntax. When the host is

unknown, use the natural-language form.

Resume routing across course modes

Before generic onboarding, resolve every "resume" or "continue" request

against these supported state files and their route owners:

• LEARNING.md belongs to learn for the full curriculum.
• MCP-LEARNING.md belongs to learn-mcp for the Model Context Protocol

(MCP) route.

• MCP-ENGINEERING-LEARNING.md is the legacy filename for that same

learn-mcp route, not a separate route.

• AGENT-SKILLS-LEARNING.md belongs to learn-agent-skills.
• CLAUDE-CERTIFICATION.md belongs to claude-certification.

If the learner names a route in a resume or continue request, dispatch to its

owner immediately even when other state files exist, then stop this skill.

For an unnamed resume or continue request, collect the owners whose state files

exist, grouping both MCP filenames under learn-mcp. If exactly one route owner

remains, invoke it and stop this skill before generic onboarding. learn-mcp

owns legacy-file migration and collision reporting. If two or more route owners

remain, list their learner-facing route names and ask which route to resume

before running placement or changing any state. If none exist, continue with

generic onboarding. Never infer a route from file recency or merge one route's

progress into another state file.

Legacy runtimes may expose learn-mcp-engineering as an alias. Accept it only

to reach learn-mcp; render every learner-facing handoff as learn-mcp and

name the route Model Context Protocol (MCP).

Focused MCP handoff

If the learner explicitly wants Model Context Protocol (MCP) rather than the

full course, do not run placement and do not create LEARNING.md. Route to

the portable skill learn-mcp, whose source is

learning-paths/model-context-protocol.json and whose state file is

MCP-LEARNING.md. Use learn-mcp in Codex,

/learn-mcp in Claude Code, or ask another compatible host to use

learn-mcp. The dedicated tutor owns lesson selection, wire

evidence, and the public-deployment security gate.

Focused Agent Skills handoff

If the learner explicitly wants Agent Skills instead of the full course, or

AGENT-SKILLS-LEARNING.md exists and they ask to resume that route, do not run

placement and do not create LEARNING.md. Route to the portable skill

learn-agent-skills, whose source is learning-paths/agent-skills.json and

whose state file is AGENT-SKILLS-LEARNING.md. Use learn-agent-skills in

Codex, /learn-agent-skills in Claude Code, or ask another compatible host to

use learn-agent-skills. The dedicated tutor owns the five-lesson order,

real-host evidence, sandbox boundaries, the Lesson 25 and tool-poisoning

prerequisite gate before Lesson 26, and the release gate.

If LEARNING.md already exists, do not overwrite it. Summarize what it says

(mission, entry point, progress so far) and offer exactly three paths:

• Resume: invoke learn with the host syntax above; skip the interview and

placement entirely.

• Re-run placement: administer the quiz again, then update only the

Placement section and the Path statuses; keep the Mission, the Progress

log, and the Review queue untouched.

• Start over: only after an explicit confirmation, rename the current

file to LEARNING-<YYYY-MM-DD>.md as an archive, then proceed with the

full onboarding below. Never delete or overwrite their history silently.

Step 1: The interview (3 questions, keep it short)

1. Why are you learning AI engineering? Free text. Examples to offer:

ship an AI product, career change, understand what I already use daily,

research. Capture their answer in their own words because it grounds every

future lesson explanation.

1. How much time per week? Options: ~2 h, ~5 h, ~10 h, "as fast as

possible". Used only to phrase the pace honestly, never to cut content.

1. What do you most want to build by the end? One line. An agent, a

trained model, a RAG product, "not sure yet" is fine.

Do not ask more than these three. The placement quiz measures knowledge;

the interview only captures intent.

Step 2: Placement

Run the placement quiz from the find-your-level skill (it installs

alongside this one): 5 areas, 10 questions, mapped to an entry phase. Preserve

that skill's answer-isolation contract: do not preload later answer-key rounds

or replace neutral <letter> placeholders with real option letters.

If the learner says they already know where they want to start ("just start

me at phase 7"), respect that and skip the quiz, with the same output

contract as a quiz run so the learn tutor always finds a well-formed plan:

• Validate the phase is 0-19 and resolve its canonical name; if it does not

resolve, list the 20 phases and ask them to pick.

• In the Path table: phases below the entry point are Skip, the entry

point and everything above are Do (no Review rows because there are no area

scores to infer them from), and the Est. hours total is the sum of the

Do rows.

• In the Placement section write Score: self-selected instead of a

number.

Step 3: Write LEARNING.md

Create LEARNING.md in the current directory with exactly these sections:

# My AI Engineering Path
<!-- Managed by the ai-engineering-from-scratch learning skills.
     Repo: https://github.com/rohitg00/ai-engineering-from-scratch -->

## Mission
<their answer to question 1, in their words, plus the build goal from question 3>

## Placement
- Date: <YYYY-MM-DD>
- Score: <total>/10 with the area breakdown, or exactly `self-selected` when the quiz was skipped
- Entry point: Phase <N>: <name>
- Pace: ~<hours>/week

## Path
| Phase | Name | Status | Est. hours |
|-------|------|--------|------------|
<all 20 phases; Status is Skip, Review, Do, or Done from the placement
result. Hours come from ROADMAP.md: read it locally if the repo is cloned,
otherwise fetch
https://raw.githubusercontent.com/rohitg00/ai-engineering-from-scratch/main/ROADMAP.md>

## Progress log
| Date | Lesson | Quiz | Note |
|------|--------|------|------|

## Review queue
<empty for now; learn adds lessons the quizzes flag>

Step 4: Hand off

Close with three lines, nothing more:

• Their entry point and total estimated hours for the Review + Do phases.
• Give the host-correct invocation for learn and say that it starts the first

lesson and picks up from this file every time.

• Give the host-correct invocation for course-guide <topic> and say that it

can jump to a specific topic instead.