five-whyslisted

# Five Whys Behind every seemingly human error is a system that allowed it. Five Whys walks the causal chain from symptom to system root cause, then applies Ries's proportional investment rule: a little fix at each level, a bigger fix at deeper levels. The classic outcome: teams stop re-fixing the same incident forever. ## The Method Ask "why?" iteratively. Each answer becomes the subject of the next why. Human error is never an acceptable stopping point — if a person could make that mistake, the system asked for it. The chain typically bottoms out at a process, a missing check, or a trade-off someone made consciously. ## Modes ### Run mode (default — user describes an incident) 1. Read `references/blame-patterns.md` before asking the first why. 2. State the symptom as observed fact, no interpretation: what happened, when, blast radius. 3. Walk the whys one at a time. Each answer must be causal ("X because Y"), not blameful ("because Dave forgot"). Struck answers get re-asked until they are causal. 4. Stop when the chain reaches a system cause that, if fixed, makes the entire chain impossible — usually between three and six whys. Five is a guideline, not a quota; say so if the chain ends early or runs long. 5. For each level in the chain, assign a proportional fix: trivial for shallow levels, structural for the root. A chain of whys with a fix only at the bottom will recur. ### Facilitate mode (user is running the analysis live with a team) 1. Copy `template