RunbookGuard

RunbookGuard is a web app (with optional CLI) that turns operational runbooks into continuously validated assets. Teams rely on runbooks for incident response, deployments, and routine maintenance, but they rot quickly: steps change, permissions drift, screenshots lie, and nobody notices until an outage. RunbookGuard connects to your infrastructure (Kubernetes, AWS, GCP, Azure) and your tooling (PagerDuty, Jira, Confluence/Notion, Git) to automatically verify that runbook steps still work. It does this by executing safe “dry-run” checks or sandboxed commands on a schedule, flagging broken steps, missing permissions, outdated links, and mismatched service names. It generates a clear “confidence score” per runbook, a changelog of what broke and when, and suggested edits. The AI component helps convert messy docs into structured steps, proposes fixes, and drafts pull requests back to your docs repo. This is not a generic documentation tool; it’s a runbook reliability system. The MVP focuses on one doc source, one infra target, and a small set of validations that catch the most common failures.