Resources

What to read and watch beyond the course. No "100 links" — only what has been verified.

Official documentation

LLM fundamentals

• Anthropic, "Building effective agents" — about loops and tools, no magic.
• OpenAI Cookbook — practical prompting recipes.
• "Attention is all you need" (Vaswani et al.) — if you want to understand the architecture.
• Andrej Karpathy, "Intro to Large Language Models" (video) — the best explanation from scratch.

Process engineering

• Google, "Code Review Developer Guide" — the classic of review, applicable to agent diffs.
• "Working Effectively with Legacy Code" (Michael Feathers) — about touching code without tests; useful before migrations.
• Trunk-based development — why small commits matter.
• 12 Factor App — configuration and secrets.

Security

• OWASP Top 10 — the mandatory minimum when reviewing.
• Supply chain attacks (typosquatting, dependency confusion) — why a new dependency from the agent requires checking.

Practice

• Your own AGENTS.md in every project.
• One experiment from module 01 once a month: intuition about models goes stale fast.
• Keep docs/decisions.md — in six months it will be the only source of "why it is like this".

What NOT to read at the start

• Lists of "100 prompts that will change everything" — they contradict the content of the course.
• Articles "we replaced the team with AI" — marketing, not methodology.
• Guides to specific prompting "cheat codes" — they go stale faster than the models.