What problem does it solve?
Writing Nightingale (n9e) alert self-healing scripts is error-prone: users confuse the stdin payload format with notification template syntax, misconfigure timeout and batch fields, expect recovery events that never fire, and risk running destructive commands on production hosts. This Skill produces correct, guarded task_tpl scripts and diagnoses common failures.
Core Features & Use Cases
- Script generation with correct stdin handling: Produces shell, Python, Go, or PowerShell scripts that parse the flat JSON stdin payload (event labels plus alert_severity, alert_trigger_value, is_recovered) injected by the ibex callback.
- Field semantics and validation guidance: Explains timeout defaults and limits, batch/tolerance/pause semantics, and CleanFields constraints so saved templates pass validation.
- Safety guardrails: Refuses blacklisted destructive commands (rm -rf /, shutdown, dd to disk) and wraps graylist commands (systemctl restart, find -delete, docker prune) with locks, dry-run switches, and before/after state output.
- Troubleshooting playbooks: Diagnoses stuck tasks, missing stdin labels caused by PromQL by-clause aggregation, and the misconception that is_recovered can trigger recovery actions.
- Use Case: A user asks for a self-healing script that cleans logs older than 7 days when disk usage exceeds 90%; the Skill outputs a guarded bash script with stdin parsing, a 120-second timeout recommendation, and rollback notes.
Quick Start
Ask the assistant to write a Nightingale self-healing script that restarts a service when its port check alert fires, including stdin parsing and a cooldown guard.
