Anthropic's prompt engineering docs: when to prompt-engineer and where to start

CodeByPoonam · x · 2026-09-10

Anthropic's Claude Platform docs lay out a complete prompt engineering guide: before starting, define success criteria, set up ways to test empirically, and draft an initial prompt. Not every failing eval is best solved by prompting—sometimes switching models improves latency and cost more easily.

All techniques (clarity, examples, XML structuring, role prompting, thinking, prompt chaining) live in the Prompting best practices reference, with an example-filled interactive GitHub tutorial and a lighter Google Sheets version.

Original post →

More from coding & agent

coding & agent channel →