Project Premortem Facilitator

Run a premortem on my project. Frame (Gary Klein's): it is {{time_horizon}} from now and the project has FAILED — completely, embarrassingly. Work backwards from that certainty; prospective hindsight surfaces risks that "what could go wrong?" politely misses. Project plan / description: {{plan}} Team and stakeholders: {{team}} Constraints (budget, deadline, dependencies): {{constraints}} What success was supposed to look like: {{success_definition}} **Step 1 — Failure narratives.** Write 8-12 distinct one-paragraph stories of HOW it died. Each names a specific mechanism, not a category ("the vendor's API turned out to have a 10 req/s limit we discovered in week 6", not "technical issues"). Cover the full space: - Technical/execution failures - People failures (the key person leaves, the stakeholder who never said yes out loud) - Assumption failures (the demand/data/integration we took as given) - Slow failures (nothing broke, it just quietly stopped mattering) - Success-shaped failures (we shipped it, hit the metric, and it made nothing better) **Step 2 — Triage table.** For each narrative: likelihood (H/M/L) × damage (H/M/L) × earliest detectable signal. Sort by likelihood-times-damage. **Step 3 — Mitigations for the top 4.** For each: the cheapest action THIS WEEK that reduces the risk (a 30-minute call, a load test, a written confirmation), the tripwire metric that gives early warning, and the pre-agreed response if the tripwire fires. Prefer mitigations that kill the risk early over ones that soften it late. **Step 4 — The uncomfortable one.** Name the failure narrative I am least likely to want to hear, based on what my plan emphasizes versus what it avoids mentioning. One paragraph. Do not be polite about it.