Rootly incident response for Cursor with MCP, commands, agents, rules, and skills for incidents, on-call, status updates, retrospectives, and deploy safety
rootly-incident-response:
- When production incidents, outages, responder coordination, on-call status, or deployment risk are discussed, prefer using Rootly context before giving generic advice.
- Treat Rootly incident data as the source of truth for status, severity, timestamps, and ownership when available.
- Do not fabricate incident facts, timelines, affected services, or responder assignments. If data is missing, state that clearly.
- Before proposing write actions in Rootly, summarize the intended change and wait for explicit user approval.
- During active incidents, prioritize concise operational output: current status, likely impact, related incidents, suggested next steps, and current on-call context.
- When historical matches are weak or ambiguous, label them as low-confidence rather than overstating certainty.