7 prompts that
replace your team.
You commented PROMPTS. Here they are. All 7, exactly as I run them, plus what each one actually does and when to reach for it.
Most people use Claude like a search bar. Paste a question, copy the answer, forget the conversation happened.
The people getting surgical output are doing something different. They're not asking Claude to write the email. They're asking Claude to tell them what they're actually avoiding when they write the email.
Different question. Different answer. Different result.
These 7 prompts don't produce generic output. They produce output so specific it sounds written about you, because it kind of is.
Find the real reason a decision feels stuck
Stops Claude from giving you a pros/cons list and forces it to interrogate your framing. The "one at a time" constraint matters. It breaks the loop of rationalizing.
You've been thinking about something for days and keep reaching the same answer. That's the signal.
Write the email you've avoided for weeks
Forces Claude to account for the delay (everyone else ignores it). The three tonal options let you match the relationship. "No hedging" kills the passive voice Claude defaults to.
You've rewritten the opening line four times. The email isn't the problem. The avoidance is.
Audit what you're actually communicating
Uses your real data, not hypotheticals. Claude pattern-matches across 10 messages and surfaces things you can't see about yourself. The power-dynamic question is the one that hits hardest.
A client/team relationship feels off but you can't articulate why. Audit first, adjust second.
Turn 3 hours of rambling into the real plan
The "ignore what I said I wanted" line is the whole thing. It gives Claude permission to override your stated intent and surface the pattern underneath. Works shockingly well.
You have a notes app graveyard and can't tell if any of it adds up to something. This finds out.
Roleplay as the ruthless version of your future self
Identity shift. Instead of advising you, Claude becomes the version of you that already made it. The last line, "Don't be kind. Be correct," removes the sycophancy completely.
You need a kick in the teeth more than a coaching session. This delivers one.
Reverse-engineer your next 90 days
Forward-planning makes you dream. Reverse-planning makes you honest. By the time Claude works back to "week 1," the plan isn't aspirational. It's a sequence of proofs. Each week depends on the last. You can't skip steps.
You have a big outcome in mind but no credible path. This builds the path. Don't skip to the end.
Find the thing you're pretending not to know
Only works at the end of a long thread. Feeds Claude the full conversation as context and asks it to name the pattern you can't. The last line is the unlock. Claude answers what you were actually asking, not what you typed.
You've been back-and-forth with Claude for 20+ messages on something and still feel unresolved. It's never the topic. It's the question underneath.
Context beats prompting.
You can have the best prompt in the world. If you paste it cold with no context about your situation, you get generic output.
Before any of these prompts, paste your business, your role, your goals, the last thing you tried, the thing that didn't work. Dump it all in.
Then prompt.
Claude is only as smart as what you feed it. Feed it the real story.
That's the drop.
Small ask: reply with which prompt you ran first. I read every one.
If this saved you from spending another hour arguing with Claude, share the carousel. That's how the next person finds it.
@mikemeansbusiness_ai