AI agents are intelligent assistants that understand your infrastructure and can execute tasks across your tools and systems. Agents Overview Grid

What agents can do

Infrastructure tasks:
  • Deploy applications and manage services
  • Scale resources up or down based on demand
  • Monitor system health and performance
  • Troubleshoot issues and run diagnostics
Development tasks:
  • Create and merge pull requests
  • Run tests and CI/CD pipelines
  • Manage repositories and branches
  • Generate and review code
Operations tasks:
  • Manage user access and permissions
  • Handle incident response workflows
  • Create and update documentation
  • Coordinate team notifications

How agents work

  1. You describe what you want in natural language
  2. Agent understands the context of your infrastructure
  3. Agent selects the right tools from your connected integrations
  4. Agent executes safely with confirmations for destructive changes
  5. Agent reports back with results and suggestions

Agent types

General Purpose Agent - Handles most common infrastructure and development tasks. Specialized Agents - Focused on specific domains:
  • DevOps Agent - Deployments, monitoring, infrastructure
  • Security Agent - Compliance, vulnerability scans, access control
  • Data Agent - Database queries, ETL, analytics
  • Support Agent - Troubleshooting, documentation, user help

Creating Custom Agents

The composer provides multiple ways to create and configure agents for your specific needs: AI-Generated Agents: Describe your requirements in natural language and let AI create a complete agent configuration with tools, instructions, and environment setup. Import from Presets: Load existing agent configurations from GitHub repositories, paste JSON/YAML configurations directly, or use community-created agent templates. On-Demand Configuration: Customize AI engines, select models (Claude, GPT-4o), choose execution runners, and configure specific tool sets. Agent Configuration Learn more: See the Agent Configuration Guide for detailed steps on creating custom agents.

Agent Management

Add tools: Connect new integrations to expand what agents can do. Set permissions: Control which resources agents can access. Create workflows: Save successful task patterns as reusable automation. Monitor usage: Track agent performance and optimize configurations.

Safety features

  • Confirmation prompts before destructive operations
  • Permission boundaries - agents can’t exceed your access rights
  • Audit logs of all agent actions
  • Rollback capabilities for supported operations

New to agents? Start by asking simple questions like “What services are running?” or “Show me recent deployments”.