conduct-a-tabletop-exercise

# conduct-a-tabletop-exercise Test decisions and coordination, not participants' ability to guess the facilitator's preferred answer. ## Preconditions - Confirm exercise owner, objectives, participants, scope, safety boundaries, and no-fault learning agreement. - Separate the fictional exercise channel from real incident channels and define how a real event will interrupt the exercise. ## Procedure 1. Restate objectives, roles, assumptions, timebox, confidentiality, and stop signal. 2. Establish the starting system state and evidence participants would actually have. 3. Deliver injects in sequence; give timestamps and avoid retroactively changing facts. 4. Ask participants to name decisions, authority, information needs, communications, and expected evidence. 5. Record actions, assumptions, dependencies, unresolved questions, and elapsed decision time. 6. Let plans fail where the scenario supports it; do not rescue participants with hidden knowledge. 7. Pause for safety, real-world confusion, sensitive disclosure, or an exercise-design defect. 8. Debrief objectives first, then strengths, gaps, surprises, and disagreements. 9. Convert findings into owned actions with due dates and verifiable completion criteria. 10. Schedule targeted retests for consequential gaps and preserve the exercise record. ## Guardrails - Never inject actions into production or impersonate real emergency authorities. - Do not use the exercise for performance discipline or public blame. - Mark all fictional artifacts clearly. - A lively discussion is not evidence that a control works; distinguish claimed, documented, and demonstrated capability. ## Done - Objectives and participant decisions are traceable to scenario evidence - Gaps distinguish plan, authority, people, technology, and communication failures - Actions have owners, dates, and acceptance tests - Critical remediations receive a scheduled retest