AI Playground

Essays, prompt-engineering challenges, and small front-end experiments — built and refined in dialogue with large language models.

The goal is clear prompts + readable code, not "vibe coding".

---

Live site

GitHub Pages: https://raven2cz.github.io/ai/

Three sections:

• Articles — long-form essay series on working with AI systemically (AI Effectively: A Systems Approach — Part 1 & Part 2).
• Challenges — real prompts that went wrong, then the better conversation that fixed them.
• Examples — small, self-contained demos (e.g. analog clock).

---

Repository layout

ai/
├─ .github/
│  └─ workflows/
│     └─ deploy-pages.yml        # GitHub Pages workflow
├─ data/
│  ├─ articles/<series>/<slug>/  # essay sources (article.html + meta.json)
│  └─ challenges/<slug>/         # challenge sources (problem.md, sessions, solutions)
├─ scripts/
│  └─ build.py                   # builds site/ from data/
├─ site/
│  ├─ index.html                 # landing page (with marker regions for build script)
│  ├─ articles/                  # built articles (one index.html per part)
│  ├─ challenges/                # built challenge pages
│  └─ examples/analog-clock/     # static examples
└─ README.md

Build

python3 scripts/build.py

Reads from data/, writes to site/challenges/, site/articles/ and updates site/index.html card lists between marker comments:

• <!-- ARTICLES:START --> ... <!-- ARTICLES:END -->
• <!-- CHALLENGES:START --> ... <!-- CHALLENGES:END -->
• <!-- EXAMPLES:START --> ... <!-- EXAMPLES:END -->

Data contracts

See AGENTS.md for how to add challenges and article series.