Prompting mistakes
Fix generic answers, replies that miss your question, easy agreement, and prompts that bundle unrelated jobs.
-
Why Does Claude Answer the Story Instead of Your Question? Usually because the story is specific and the question isn't. Level 1: Basic prompts
- 1Vendor missed two datesfact
- 2Launch is on March 3fact
- 3Legal hasn't approvedfact
- 4Three-step plan to decidethe ask: change vendors or not, by Friday
-
How Do You Stop Claude From Just Agreeing With You? Don't show which answer you're hoping for. Level 2: ContextShows your preferenceHides your preference
- "Isn't this right?"
- "I lean A, thoughts?"
- Options A and B
- Fresh chat, criteria first
-
Should You Split a Big Claude Prompt Into Steps? Only when you need to check the result between steps, or when the steps must run in a fixed order. Level 1: Basic promptscost of a wrong first step →Separate chatscheck each on its ownNumbered list, then a stopcheck before the rest runsNumbered lista missing part is easy to spotOne promptClaude handles most steps internallysteps depend on each other →
-
Why Does Claude Give Generic Answers? Usually because the prompt left things for Claude to guess. Level 1: Basic prompts
- "Fix the login bug"no symptom, no place, no proof
- Add the symptomfails after session timeout
- Add where to looktoken refresh
- Add the proofa failing test, then the fix
How to tell which mistake you made
Look at the answer you got. If it's correct but could apply to anyone, the prompt was vague. If it discusses your situation but gives you nothing to use, the request was missing or unclear. If some parts are strong and others are thin or absent, you bundled too much. If it praises your idea, you may have shown which answer you wanted.
Rewrite the prompt, not the answer
When an answer misses, a follow-up correction feels quicker, but each one adds to the conversation. Anthropic's Claude Code guide says that after two failed corrections, you should clear the session and write a better prompt. Rule of thumb, not a vendor figure: use the same approach in chat when your second correction hasn't fixed it.
Why are Claude's answers so generic?
The prompt probably left out the reader, purpose or standard, so Claude wrote for a typical case. Anthropic's docs say the more precisely you explain what you want, the better the result.
Why does Claude agree with everything I say?
AI assistants can drift toward what a user seems to want, as Anthropic's research on sycophancy describes. Present options neutrally, set your criteria first, and turn memory off for that chat from the + menu.
Should I put several tasks in one Claude prompt?
You can if they're related and listed clearly. Anthropic's docs recommend numbered steps when order or completeness matters. Split them into separate turns when you need to check one result before the next step.
How long should a Claude prompt be?
As long as it needs to be to cover the reader, purpose, facts and standard. Length itself doesn't help. A short prompt with clear limits usually does better than a long one with none.