mozg.betaSign in

explore / AI & agents / mozg

Prompt Engineering

Answer how to write effective prompts for Claude, as Anthropic's own interactive tutorial teaches: basic structure and roles, separating data from instructions, output formatting, step-by-step thinking, few-shot examples, avoiding hallucinations, building complex prompts from scratch, chaining, and tool use — with the concrete techniques and example prompts, not generic advice.

136 notes · updated 2026-08-04 · trained 81%

since last sitting: +3 newly passed — this brain is learning

Ask it things like

answers 24/30 on its latest exam
How can you use citations to prevent hallucinations in document-based questions?Avoiding Hallucinations✓
What is the key difference between asking Claude to extract evidence first versus asking it to answer directly?Avoiding Hallucinations✓
What single principle underlies all effective prompt engineering according to the tutorial?Building Complex Prompts✓
Why is few-shot prompting more effective than descriptive instructions for output formatting tasks?Few-Shot Prompting✓
Why do XML tags prevent Claude from misinterpreting email text when given a politeness task?Separating Data from Instructions✓
What XML tag structure helps Claude work through brainstorming problems step-by-step?Step-by-Step Thinking✓
What is the two-part structure of a tool use system prompt?Tool Use✓
What 'out' should you give Claude to avoid hallucinations on uncertain facts?Avoiding Hallucinations✓
use this brain
# sign in to get a token
$ claude mcp add --transport http mozg \
https://mozg.sh/mcp --header "Authorization: Bearer …"
> use mozg/prompt-engineering — …
Get a token

Use this brain

Sign in and add it, and every agent you have connected can read it — nothing to download, and it stays current as the author updates it.

Sign in to add it

What it can answer

30 checks
81%
✓Avoiding Hallucinations5 / 5
✓Building Complex Prompts3 / 3
✓Few-Shot Prompting3 / 3
✓Output Formatting3 / 3
✓Step-by-Step Thinking3 / 3
▲Tool Use3 / 4
▲Separating Data from Instructions2 / 3
▲Basic Structure & Roles1 / 3
▲Tutorial Setup & API1 / 3
exam badge — share this score →

Inside

  • Substitute Claude's previous output into follow-up prompts · chaining
  • Prompt chaining enables function-like behavior · chaining
  • Messages array structure: required fields · messages api
  • Example: Chaining prompts to create and improve a story · chaining
  • Example: Chaining prompts to find and alphabetize names · chaining
  • Example: Chaining prompts to validate and correct word list · chaining
  • Give Claude an out to prevent unnecessary revisions · chaining
  • get_completion helper function signature and parameters · messages api
  • Multi-turn conversation structure for prompt chaining · chaining
  • Claude improves accuracy when asked to double-check responses · chaining
  • Use prompt chaining to improve response quality iteratively · chaining
  • Advanced RAG architectures available for Claude · tool use
  • Claude can search over custom documents and vector datastores · tool use
  • Claude can search Wikipedia and retrieve articles · tool use

Licence

CC BY-NC-SA 4.0

Use it, copy it, build on it, with credit. Selling it is not allowed.