What problem does it solve? Operational procedures often live only in one engineer's head, so when something breaks at 3am nobody knows the exact commands to run. This Skill captures deploys, rollbacks, alert responses, and disaster recovery steps as verified, copy-pasteable runbooks indexed by symptom. ## Core Features & Use Cases - Runbook Authoring: Generates short, step-by-step runbooks in docs/runbooks/ with exact commands, verification lines after destructive steps, named escalation paths, and rollback-of-the-rollback instructions. - Symptom-First Index: Maintains docs/runbooks/INDEX.md keyed by observable symptom rather than system name, so only the index and one relevant runbook are loaded into context. - Verification Tracking: Stamps every runbook with last verified metadata and audits entries not verified in 90 days or referencing resources that no longer exist. - Use Case: After resolving an incident where inventory files stopped loading, ask the Skill to draft a rollback runbook from the actual commands run in the session, then add it to the symptom index marked last verified: NEVER until someone executes it end to end. ## Quick Start Ask the assistant to write a runbook for rolling back the inventory loader Lambda, including verification steps and an escalation contact.
