What problem does it solve? Agent sessions produce work but often lose the lessons learned, forcing the next agent to repeat the same trial-and-error. This Skill enforces a write-back loop so every non-trivial session ends with updated skill files, verified examples, and filed issues for unresolved gaps. ## Core Features & Use Cases - Failure Triage: Checks failed Argo Workflows from the past 7 days at session start and routes each failure pattern to the relevant skill file. - Skill Routing Table: Maps changed areas (Argo Workflows, KubeVirt, ArgoCD, tests, CI) to the specific skill file that must be updated. - Documentation Hygiene: Verifies code examples against Context7 docs, audits skill files against the canonical spec, and removes dated or resolved entries. - Use Case: After debugging a KubeVirt VM provisioning failure, route the discovered pattern into kubevirt-vms/SKILL.md, verify any Argo examples against Context7, and file a GitHub issue for remaining infra gaps before creating the PR. ## Quick Start Run the skill improvement loop to check recent nightly failures and write back any patterns discovered in this session to the appropriate skill files.
