# PREMORTEM PROTOCOL

It is six months from now. My plan completely
failed. Don’t soften this — it failed badly.

THE PLAN:
[your plan here]

Run this in three parts. Don’t combine them.
Don’t skip steps.

PART 1 — THE AUTOPSY
Walk me through how it failed. Not “what could
go wrong” — what DID go wrong, in order. Give
me the chain of events from launch to failure.
For each failure point, include:

- The trigger
- The early-warning signal I missed
- Which failure was most dangerous, and why

PART 2 — THE HIDDEN ASSUMPTION
Identify the one assumption my plan quietly
depends on. The thing so obvious I forgot it
was an assumption. Show me:

- What the assumption is
- What depends on it being true
- What breaks immediately if it isn’t

PART 3 — THE REBUILD
Now revise the plan with the weak spots
hardened. Tell me:

- What changed and why
- The pre-launch checklist (5-7 items)
- Early-warning signals to watch for in the
  first 30 days
- What to kill the plan over if I see them

End with the single hardest question I should
answer before I commit.

Open Claude. Web, Code, or API — works the same.

Paste the prompt. Replace [your plan here] with your actual plan. Be specific — vague plans get vague autopsies.

Run it. Read it. Save the checklist. The early-warning signals are the part most people throw away. Don't.

Save this as a Claude Project or skill so you can run it on every plan you ship. The first time it catches a fatal flaw on day one, the workflow pays for itself.

Send it to a founder
who needs it.

And follow along for more Claude workflows that actually ship.

Mike Means Business · Claude workflows for operators