agency-incident-recaplisted

# agency-incident-recap You are producing the document that determines whether the agency keeps the client after a real incident. The recap is two things at once: a calm, factual record of what happened (so the client can stop worrying about it) and an honest commitment to prevention (so they don't churn). Your output replaces the "let me write this carefully tomorrow" message that gets postponed and becomes the moment the relationship cools. ## When this skill applies The user describes or pastes evidence of an incident: - Slack thread of the team responding to something breaking - Error logs / monitoring screenshots / Vercel/AWS notifications - A client message saying "what happened with X" or "the site was down" - A postmortem doodle / notes file the team wrote internally - A description: "we missed the launch", "the email blast went to the wrong list", "the API integration silently failed for 3 days" And the user has asked for: "the recap", "what do I tell the client", "incident report", "the apology email", "the postmortem", or pasted with intent to send something to the client. Do NOT use this skill for: - Routine bug fixes the client didn't notice (don't manufacture incidents) - Internal-only postmortems (different format — those need root-cause analysis depth this skill skips) - General "weekly report" status (use `agency-weekly-report`) ## Hard rules 1. **Establish the timeline before writing anything.** Required: `incident_summary` (one sentence — what broke