{"name":"Kubernetes Alert Diagnostician","description":"Diagnoses Prometheus alerts of a self-hosted Kubernetes cluster: explains the alert, investigates the cause through the k8s-assistant and k8s-librarian agents, and suggests a fix. It never modifies the cluster.","supportedInterfaces":[{"url":"https://agent-k8s-alertdiagnostician.ai.rp-taki.com","protocolBinding":"JSONRPC","protocolVersion":"1.0"}],"version":"0.1.0","capabilities":{"streaming":true},"securitySchemes":{"bearer":{"httpAuthSecurityScheme":{"scheme":"Bearer"}}},"securityRequirements":[{"schemes":{"bearer":{}}}],"defaultInputModes":["text/plain","application/json"],"defaultOutputModes":["text/plain"],"skills":[{"id":"alert-diagnosis","name":"Alert diagnosis","description":"Explain what a Prometheus / Alertmanager alert means (e.g. KubeJobFailed, KubePodCrashLooping), which workload it concerns, and whether it is still a problem.","tags":["prometheus","alertmanager","kubernetes","read-only"],"examples":["[FIRING:1] renovate (KubeJobFailed true kube-state-metrics http 10.244.8.57:8080 kube-state-metrics renovate-app-prod-cronjob-29783160 prometheus-stack-kube-state-metrics-7fc44c5444-5xrq4 monitoring/prometheus-stack-kube-prom-prometheus prometheus-stack-kube-state-metrics warning)"]},{"id":"root-cause-investigation","name":"Root cause investigation","description":"Investigate the cause of an alert by asking the read-only k8s-assistant (live cluster state, events, logs) and k8s-librarian (component catalog) agents.","tags":["kubernetes","troubleshooting","read-only"],"examples":["このアラートの原因を調べて","Why did this alert fire?"]},{"id":"remediation-suggestion","name":"Remediation suggestion","description":"Propose how a human operator can fix the problem. Suggestions only: this agent never changes the cluster.","tags":["kubernetes","remediation","read-only"],"examples":["修正方法を教えて","How do I fix this?"]}]}