Skip to main content

Prompting Claude Specifically

Intermediate

Most of prompting is universal. But a few techniques are especially effective with Claude — and a couple of habits from older models are now counterproductive. This lesson separates the signal from the noise: what's actually Claude-specific.

What you'll learn
  • Recognize the Claude-specific techniques worth leaning into — XML tags, clarity, roles, thinking, examples, and prefilling
  • Unlearn the old-model habits (ALL-CAPS, threats, over-engineered scaffolding) that now backfire
  • Know which two habits — XML-tag affinity and prefilling — to carry as your Claude signature
  • Understand what to re-check when you switch models

Lean into these with Claude​

These are the moves that pay off the most with Claude. The first one is the single highest-value habit you can build.

  • XML tags. Claude is well-tuned to structured prompts with named tags for inputs, instructions, and examples. This is the highest-value Claude habit.
  • Be clear and direct. State the goal, constraints, and what "done" looks like. Claude follows explicit instructions closely.
  • Give it a role via the system prompt.
  • Let it think on hard problems — ask for step-by-step reasoning, or use extended thinking/effort on reasoning models (where you often don't need to ask for CoT explicitly).
  • Use examples (few-shot) for format and edge cases.
  • Prefill the start of Claude's response (on the API) to steer format.

Try prefilling​

Prefilling means starting the assistant turn yourself so Claude continues in the format you want. On the API, begin the assistant turn with an opening character or heading:

Prefill to push JSON output

{

Prefill a heading to push structure

## Summary
Pro tip
  • Begin the assistant turn with { to push JSON, or with a heading to push structured output. Prefilling is an API-only lever — it steers format without spending words on instructions.

Drop these old habits​

Newer, literal-following Claude models change the rules. Tricks that helped on older models now get in the way.

Watch out
  • Ease off ALL-CAPS and threat-style prompts ("YOU MUST… OR ELSE"). Newer Claude models follow instructions literally, and politely-phrased, specific instructions work better than shouting.
  • Don't over-engineer with elaborate "jailbreak-style" scaffolding — clear, honest instructions outperform tricks.
  • Less is often more — a tight prompt beats a sprawling one that buries the ask.

Claude vs other models​

The principles transfer (see Cross-AI Prompt Translation), but two things are most worth remembering as Claude-specific.

Guided walkthrough1 of 3
  1. Wrap inputs, instructions, and examples in named tags. Claude's tag affinity makes this the most reliable structuring habit.
Claude-specific quick reference
Press Enter or Space to flip the card. Use the left and right arrow keys to move between cards.Term shown.
1 / 5
Key takeaways
  • XML tags are the highest-value Claude habit; prefilling is the most useful Claude-specific format lever.
  • Newer Claude models follow instructions literally — be clear and direct, not loud or threatening.
  • Less is more: a tight, honest prompt beats elaborate scaffolding.
  • Principles transfer across models, but re-check tone intensity and formatting cues when you switch.

Check yourself

0/3
  1. Which technique is described as the highest-value Claude-specific habit?
  2. How does prefilling steer Claude's output format on the API?
  3. What should you do with ALL-CAPS and "YOU MUST… OR ELSE" prompts on newer Claude models?

Next​