autonomous-agents

SUSPICIOUS. The monitoring purpose broadly matches the DB, Redis, scheduler, and alerting capabilities, but the footprint is too broad for a normal docs/health skill: it uses root SSH to a raw server IP, reads worker env files, autonomously sends external WhatsApp messages, and embeds real-looking credentials in the skill text itself. The install path is only moderately risky, but the credential exposure and autonomous outbound alerting make the overall skill high risk.