Docs/Agent Capabilities/Business Automation

Business Automation

Real-world use cases: support, ops, monitoring, reporting, and coordination.

SkyKoi agents aren't just chatbots — they're full-featured AI assistants that can automate complex business workflows with shell access, web browsing, and scheduled tasks.

Development Assistance

  • Write, review, and debug code directly on the server
  • Run build pipelines and tests via shell commands
  • Research APIs and documentation through web search and browsing
  • Manage git repositories (clone, commit, push, review PRs)
  • Generate and edit configuration files

Infrastructure Operations

  • Monitor infrastructure health and alert on anomalies
  • Run deployment scripts and manage CI/CD pipelines via shell commands
  • Generate daily/weekly reports
  • Automate repetitive tasks with cron-scheduled agent runs
  • Manage AWS resources through natural language

Data Pipeline Management

  • Trigger and monitor ETL jobs
  • Query databases and generate reports
  • Process files (CSV, JSON, logs) with shell tools
  • Sync data between services
  • Alert on pipeline failures

Monitoring & Alerting

  • Heartbeats — The agent checks in periodically (configurable) to monitor systems, email, calendars, notifications
  • Cron jobs — Schedule exact-time checks: "Every Monday at 9 AM, review infrastructure costs and generate a report"
  • Proactive alerts — Agent can detect issues (high CPU, failed deployments, etc.) and log them

Example: Automated Ops Bot

text
# Agent SOUL.md
You are OpsBot, the infrastructure monitoring assistant for Acme Corp.

Your responsibilities:
- Monitor AWS infrastructure health every 4 hours (via heartbeat)
- Run daily cost reports at 9 AM EST (via cron)
- Deploy to staging when asked, require approval for production

Tools you use:
- exec: AWS CLI, kubectl, docker
- web_fetch: Grafana dashboards, PagerDuty API
- cron: Schedule recurring checks