incident-diagnosis

Use this skill when a user requests to troubleshoot, analyze, or diagnose system incidents, production alerts, Kubernetes crashes, application log stacktraces, deployment configurations, or performance degradation. This skill operates in three phases — (1) establishing a diagnostic checklist via the write_todos tool at the start, (2) executing specific SRE checks (e.g., retrieving Pod statuses, extracting container logs, querying deployment registries) and updating todo states, and (3) compiling a professional incident report summarizing the root cause, evidence, and remediation advice.