9 chapters + appendix, each with a hands-on "Example Playground." Anthropic already ships the interactive version (Jupyter + Google Sheets) — this is the map; the timed drills below are the zero-setup way to rep the same skills. See also the docs guide ↗ and the full prep guide §5.
4Separating data from instructions5Formatting output & speaking for Claude6Precognition (think step by step)7Using examples (few-shot)
Advanced
8Avoiding hallucinations9Building complex prompts (industry use cases)
Appendix — beyond standard prompting
Chaining promptsTool useSearch & retrieval
The drills below rep: clear & direct (2) · separating data / XML (4) · output formatting (5) · step-by-step (6) · avoiding hallucinations (8).
⚡ Timed drills — write → run → refine
Edit the prompt until every case passes — the CodeSignal write→run→refine loop.
Grading is strict (exact JSON, exact tag contents, word counts), same as a programmatic grader.
Use {{INPUT}} where the test input should go.