Agent presets
Configure presets under Agent Presets in the space sidebar. Each preset defines:| Field | Purpose |
|---|---|
| Harness | Agent runtime (for example, Claude Code, Codex) |
| Sandbox | Execution environment (Arize-managed, Daytona, Vercel, etc.) |
| Skills | Optional—GitHub, Arize, Datadog, custom |
| GitHub repo | Shown when a GitHub skill is attached |
Session vs automation
| Mode | Use for |
|---|---|
| Session | One-off work; follow the transcript and send follow-ups |
| Automation | Repeating task on cron, metric threshold, or monitor |
Start a session
Enter task prompt
Describe the investigation or output you need (for example, root-cause a tool failure, draft eval labels for faithfulness failures).
Job detail
On start, the worker provisions a sandbox, clones the repo (if configured), loads skills, and pulls project context. From job detail you can:- Read the live transcript
- Send follow-ups to steer the run
- Reopen a completed session from Fleet Observability (sandbox restarts)
Outputs
| Output | When |
|---|---|
| Investigation markdown | Always possible |
| Eval labels, dataset rows | Task-dependent |
| Branch / PR | Preset includes repo write access |
Create an automation from Studio
- Click + Setup as automation
- Set trigger: cron, metric threshold, or monitor
- Pause, edit, or delete from Fleet Observability
Learn more
These pages sit under Managed Agents in the sidebar:- How agents work — Sandboxes, data access, credentials, and security boundaries
- Skills and permissions — Configure GitHub, Arize, Datadog, and custom skills; RBAC