Challenge course · Any coding agent (Claude Code 2.x / Codex CLI 0.x / Cursor)

One source of truth for agent rules

Stop maintaining diverging AGENTS.md, CLAUDE.md, and .cursorrules files — build a checker that proves your repository has exactly one rule source.

Intermediate5 challenges1–2 hoursMaintains rule files for multiple toolsFreeCourse badge: Measure the driftReviewed 2026-09-12English + 中文

Start it with your agent — paste this one sentence:

Read https://flypython.com/skills/flypython/SKILL.md and start the FlyPython course `agent-rules-single-source`.

Needs a coding agent that can run commands and reach the network (chat-only web AIs cannot). Your agent authorizes you with a one-time link, fetches this course's files itself, and you never download anything. Haven't installed the FlyPython Skill? Pick your agent's tool course →

Browse the challenges 中文版 →

What you learnA working rule-consistency checker (rules_check.py) that fails CI the moment a rule file drifts from AGENTS.md — plus the single-source setup applied to your own repository.

Who it's forMaintains rule files for multiple tools · Python 3.11 or newer available on your PATH

Duration & cost1–2 hours · Free

Challenges

The challenges — read them online

No clone, no account — click a checkpoint to read the full challenge on its own page.

Measure the drift you already have

Run the checker against three scenario repositories and read drift as a test failure, not a vibe.

Checkpoint l01 · 10 pts · read online →02

Decide the source of truth

AGENTS.md becomes the one file agents read; everything else is a pointer or a copy.

Checkpoint l02 · 10 pts · read online →03

Drive a bounded change, test by test

Make the starter detect source, pointer, copy, and drift — one failing test group at a time.

Checkpoint l03 · 10 pts · read online →04

Verify and wire it in

Prove both verifier directions, run all three skins, then add the checker to your review loop.

Checkpoint l04 · 10 pts · read online →05

Apply this to your repository

One AGENTS.md, thin pointers everywhere else, and a failing check when it breaks.

Checkpoint l05 · 10 pts · read online →

Start

How the run works

Your agent drives; the files arrive through the Skill — no download, no clone.

SkillAny coding agent (Claude Code 2.x / Codex CLI 0.x / Cursor)

The Skill does the fetching

The sentence above starts the course: your agent walks you through the one-time authorization link (compare the code, click Allow), fetches this course's files, and drives the challenges with you. Agents without the Skill installed can work from this brief instead:

Work on the FlyPython challenge "One source of truth for agent rules" (course id course-agent-rules).
Machine-readable brief: https://flypython.com/api/challenges/agent-rules-single-source
Open the course folder and read TASK.md first — it is the contract.
Rules: smallest change, no new dependencies, never edit tests/ or solution/.
Check with python verify.py until its gates pass, then report each claim code to me.

Your agent works under the task contract: smallest change, no new dependencies, never edit solution/. Optional guided mode: COURSE.md.

Files

Files come to you

Course files live at /api/challenges/agent-rules-single-source/files — your agent fetches the manifest and writes each file at its path. There is nothing to download by hand.

How "done" is decided →

Scenario skins

Same skills, a domain you recognize

All three skins run through the same code and the same tests. Pick the one closest to your job before lesson 1.

scenario/thin-pointer/AGENTS.md + CLAUDE.md

Healthy: thin pointers

AGENTS.md plus two thin pointer files that defer to it.

scenario/single-copy/AGENTS.md + CLAUDE.md

Healthy: exact copy

AGENTS.md plus one byte-identical copy in CLAUDE.md.

scenario/drifted/.cursorrules

Broken: drifted .cursorrules

A stale full ruleset that contradicts AGENTS.md — the drift to catch.

Verification

How “done” is decided

An 11-test suite pins pointer detection, the source requirement, copy and drift classification, and CLI exit codes; the starter reproduces six named failures, the solution passes 11/11.

python verify.py  # from the course folder: checkpoint status + claim codes

Apply this to your own project

Lesson 5 ports the workflow — not the code — to one script you actually own: a three-line task contract, one new failing test made to pass, and a written record of what the tests do not prove. The same loop is described in the AI coding workflow guide and practiced by therunnable examples.

Claim your evidence

When a checkpoint passes, python verify.py prints a deterministic claim code. Your authorized agent submits them for you in one batch; by hand, paste them on yourprogress page — worth 10 points per checkpoint and theMeasure the drift course badge at five. Claim codes are self-reported evidence, not certificates.

What verification does not prove

Passing the course suite proves the pinned behaviors on the tested inputs — not correctness on tomorrow’s data, not production performance, and not anything about your own project until you write its contract. Repository tests are evidence, not your outcome.

Report teaching drift ↗All courses →