cycle-aware-verification overview

Composable cycle-aware verification component that validates fixes survive system execution cycles (cron jobs, scheduled tasks, server restarts, watch-mode rebuilds). Many bugs only manifest after a system completes its next cycle — a server that passes an immediate health check may crash on its next hot-reload, a cron fix may break on the next scheduled invocation, or a build-watch process may fail on the next file-change trigger. Traditional verification checks the system once, immediately after a change. Cycle-aware verification adds a temporal dimension: it performs a baseline check, waits for the system to complete its next execution cycle, then re-checks. A fix is only considered valid if the system survives both the immediate check and the post-cycle check. The module also provides pre-flight analysis to scan files for dangerous patterns before changes are applied, catching destructive commands (rm -rf, kill -9, etc.) before they can do damage. The module exposes four surfaces: - `createPreflightAnalysis(config)` — factory that builds a shell task to grep for dangerous patterns in a target file. - `cycleAwareVerificationTask` — standalone `defineTask` descriptor (kind: 'shell') that performs baseline check, waits for a cycle, then re-checks in a single shell invocation. - `createCycleAwareVerification(config)` — factory that returns separate baseline and post-cycle task definitions for fine-grained manual control. - `createPostCycleSurvivalCheck(config)` — factory that builds a single post-cycle survival check task with baked-in URL, cycle interval, and expected status.