Claude · Module 08 of 8

Mistakes

The common errors that make Claude output worse — and how to avoid them.

2 lessons8 exercises
Topics & techniques

Prompting mistakes

The vague-prompt traps that waste your time.

4 cards →
Mistake
Don't Bury the Question in Backstory
A common mistake of writing a long run-up — background, history, caveats, feelings about the problem — and only at the very end, almost as an afterthought, stating what you actually want. The real request gets lost in the preamble, and the answer misses the point.
L1 · Basic prompts
Mistake
Don't Lead the Witness
The mistake of phrasing a prompt so it signals the answer you're hoping for — 'This is the better option, right?', 'Don't you think X is a mistake?' — and then treating Claude's agreement as independent confirmation. You fished for a conclusion and caught the one you baited.
L2 · Context engineering
Mistake
Don't Make One Prompt Do Five Jobs
The mistake of packing several distinct tasks into a single prompt — research this, then write it up, then edit for tone, then format as a table, and also suggest a title — and expecting all of them done well at once. Some parts land, others get skimmed or dropped.
L1 · Basic prompts
Mistake
Vague Prompts Waste Turns — Specify the Target
The most common prompting mistake: asking for something under-specified ('make this better', 'write about marketing') and getting a generic, average answer.
L1 · Basic prompts

Verification mistakes

The ways people over-trust AI output — and the cheap checks that prevent it.

4 cards →
Mistake
Check Against the Source, Not the Summary
The mistake of accepting Claude's summary, quote, or paraphrase of a document as accurate because it reads well — without checking the specific claims back against the source. A fluent summary can still misstate, over-generalise, or invent a detail the original never contained.
L2 · Context engineering
Mistake
Don't Outsource the Final Judgement
The mistake of letting Claude make the call rather than inform it — shipping its output because it seems reasonable, without the review and judgement that the decision actually needs. The model is there to draft, analyse, and pressure-test; the decision and the accountability stay with you.
L2 · Context engineering
Mistake
Don't Ship Code You Only Read
The mistake of accepting AI-generated code because it looks correct on the page — well-structured, sensibly named, plausible logic — and shipping or committing it without actually running it, testing it, or exercising the edge cases. Readable is not the same as working.
L2 · Context engineering
Mistake
Don't Trust Confident Specifics Without Checking
The mistake of taking Claude's most confident-sounding specifics — exact figures, quotes, citations, API names, legal or medical claims — at face value because the fluent, assured tone reads as authority. Confidence is a property of the prose, not evidence the fact is right.
L2 · Context engineering
Common Claude mistakes

Vague prompts make Claude guess

The most common way to waste a Claude session is to under-specify. When you leave out the role, the goal, the constraints, or the example, Claude has to guess at your intent — and a good guess is still a guess. The fix is not longer prompts but clearer ones: say who Claude is, what outcome you want, and what good looks like. Precision up front saves the back-and-forth that vague prompts inevitably create.

Fluent is not the same as correct

Claude is careful and tends to stay grounded, but it is still capable of confident errors, and its polish can lull you into skipping verification. Build the habit of checking: ask 'what would break this?', request the reasoning behind a claim, and confirm anything important against a source or a test. The mistake is treating a well-written answer as a verified one. A verification step is cheap; shipping a plausible-but-wrong answer is not.

The mistake topics to explore

This section covers prompting mistakes (the vague-prompt traps) and verification (how to check Claude's work before you rely on it). Each card below pairs a common error with its fix.

Frequently asked
Why does Claude misunderstand my prompts?

Usually because the prompt left out the role, goal, constraints, or an example, forcing Claude to guess your intent. Being explicit about who Claude is and what a good answer looks like resolves most misunderstandings.

Does Claude hallucinate or make mistakes?

Claude is careful and tends to stay grounded, but it can still produce confident errors. Verify anything important — ask for reasoning, request sources, or check against a test — rather than trusting a polished answer on sight.

How do I verify Claude's answers?

Ask 'what would break this?', have it explain how it reached a claim, and confirm key facts against a source or a quick test. Treating verification as part of the workflow catches the errors that fluent writing can hide.

What is the biggest mistake when using Claude?

Trusting fluency. A well-written answer is not necessarily a correct one, and skipping verification on work that matters is where people get burned. Build a lightweight checking step into how you use it.

Test yourself

Are you making these Claude mistakes?

Six questions on the errors that quietly waste your turns — vague prompts, leading questions, and trusting fluent output you never checked. No sign-up, instant score, and you can share how you did.

  1. Q1 Why does 'make this better' reliably produce a mediocre answer?
  2. Q2 You ask 'This is the better option, right?' and Claude agrees. What have you learned?
  3. Q3 You write three paragraphs of background and end with 'thoughts?'. What went wrong?
  4. Q4 You pack research, writing, editing, formatting, and a title suggestion into one prompt. What happens?
  5. Q5 Claude's summary of a document reads clearly and confidently. What does that tell you about its accuracy?
  6. Q6 AI-generated code is well-structured, sensibly named, and reads correctly. Ship it?
0 / 6 answered