What problem does it solve? Reviewing a mature homelab by hand means re-checking dozens of settled decisions, duplicating findings across domains, and shipping unverified recommendations. This Skill coordinates a multi-agent audit of the entire k3s/Ansible homelab, deduplicates and adversarially verifies findings, and produces one prioritized report without changing any infrastructure. ## Core Features & Use Cases - Parallel per-domain review: Dispatches six read-only reviewer agents covering security, network/reverse proxy, backups/observability, CI/CD, container infra, and docs-vs-config drift, each with pinned model sizing. - Memory-primed deduplication: Reads standing do-not-reflag lists and open-finding registers first, so settled trade-offs are not re-flagged and recurring findings are reported as recurrences rather than discoveries. - Adversarial verification: Runs skeptic agents against every High/Medium finding and every proposed remediation, confirming, refuting, or flagging fixes that launder symptoms instead of resolving defects. - Use Case: Ask for a security and backups review of the homelab; the Skill primes from prior review ledgers, dispatches the two relevant reviewers in parallel, verifies their findings against cited file:line evidence, and files confirmed gaps as GitHub Issues. ## Quick Start Ask the assistant to run a homelab review of the security and backups areas and report prioritized findings without making any changes.
