GV

GeezerVison

The World Wide Web, restored to 800 pixels of glory.

Home|About|Gaming cabinet|Archives|How it works|Starter sites

On this site

  • FlyPython
  • Courses
  • Tracks
  • Leaderboard
  • Learn
  • Radar
  • Pricing
  • 中文本页暂无中文版
  • Sign in
  • Examples
  • Resources
  • Badges
  • Search
  • Help

This view

  • Original page
  • New address

GeezerVison is showing flypython.com — FlyPython — Learn to ship Python with AI coding agents

Real projects · verify.py decides when you're done

Learn to ship Python with AI coding agents.

Real projects. Your agent does the typing; verify.py decides when you're done. Coach P is your AI coach inside your own agent, using your model: first a question, then a small hint. Pick your tool — lesson 1 installs the FlyPython Skill for you.

Start with your agent's tool — one click, no sentence to copy:

Claude Code 2.xStart course →OpenAI CodexStart course →CursorStart course →DeepSeek HarnessStart course →Kimi CodeStart course →ZCodeStart course →GitHub Copilot in VS CodeStart course →

Every tool course is free. Inside, one sentence starts the course in your agent — it authorizes you, fetches the files, and you never download anything.

Open source18 courses · 94 claimable challengesFree challenges include all seven agent-tool coursesEnglish + 中文

What is the FlyPython Skill? →Browse tracks →All courses →
A place to startPython 3.11+

A broken script.
A working report.

Hands-on Python with Claude Code

Start
A script that crashes on real data
Build
A tested report tool
Check
Five local verification checkpoints
Open the challenge →

How it works

Authorize. Solve. Claim.

Every challenge ships a task contract and a local verifier. Nothing is graded on vibes.

01 · Authorize

One link, no passwords

Your agent shows you a link and a short code. You log in, compare the code, click Allow — the agent gets a scoped token. It then fetches the course files itself.

02 · Solve

You drive, the agent works

It reads the contract, edits the starter, and runs python verify.py at every gate — asking you the checkpoint questions before claiming anything.

03 · Claim

Evidence, self-reported

The verifier prints the claim codes; your agent submits them in one batch. Points, badges, and rank record themselves — honest evidence, never a certificate.

Guides →Playbooks →Examples →How we publish →

Tracks

Structured routes, real badges.

Sequenced courses, a capstone, and a badge when every module is claimed.

11

Agent tools foundation

Learn to drive real coding agents — Claude Code, Codex, Cursor, GitHub Copilot, DeepSeek Harness, Kimi Code, or ZCode — give them good rules, and verify everything they produce.

1110 pts · Badge: Agent User →05

Data analysis with agents

From a messy export to a verified analysis report: clean, explore, visualize, and report on a real dataset with your agent.

600 pts · Badge: Data Analysis Agent →04

Python developer toolchain with agentsPro

Ship like an engineer: a real pytest suite that catches seeded bugs, a packaged CLI, and a verified web API — built with your agent, proven by verify.py.

500 pts · Badge: Toolchain Builder →

Meet the builders
on the leaderboard.

Practice, complete challenges, and earn your place. These are the top five builders by recorded points.

See the full leaderboard →Last 30 days →By course →

Points reflect self-reported challenge claims.

BuilderPoints

Open the full leaderboard to see current standings.

    You are visitor 01402503 — GeezerVison does not store the pages you look up.
    Best viewed with Netscape Navigator 4.7 or Internet Explorer 6 at 800×600. No cookies required.