Pylon Agent Configuration Reference
Reference for the Pylon Agent editor - Overview, Tools, Skills, Sandbox, Limits, Triggers, Config, and Runs tabs, with fields, defaults, and behavior.
Purpose
Section titled “Purpose”Every tab of the Agent editor, with what each field does. To build your first Agent end to end, see Create your first Agent; to put one on a cadence, see Schedule an Agent.
The editor tabs
Section titled “The editor tabs”| Tab | What it controls |
|---|---|
| Overview | Identity and brains: Name, Description, Model (picked from your account’s chat-model catalog), System prompt |
| Tools | What the agent can do per turn: built-in tools (always available, read-only) plus any MCP servers you attach |
| Skills | Instruction packs: toggle catalog skills on; the agent loads their instructions before acting |
| Sandbox | Read-only facts about the isolated environment each run gets |
| Limits | Per-run guardrails (table below) |
| Triggers | What starts the agent: Manual, Schedule, GitHub webhook, Agent-to-agent - each addable once |
| Config | The canonical configuration document (below) |
| Runs | Run history for this agent - still rolling out; use Monitor → Activity meanwhile |
Tools: built-in vs attached
Section titled “Tools: built-in vs attached”The built-in tool set is always available and can’t be edited; it’s listed on the Tools tab under “Built-in tools - always available”, grouped as Built-in, Coding, Approval-gated, Planning, Forms, and Chat. The one way to extend an agent’s tools is attaching an MCP server - pick an existing one or Register a server with a name and optional description. Manage the servers themselves under Deploy → Capabilities → MCP Servers.
Sandbox
Section titled “Sandbox”Each run executes in an isolated micro-VM created for that run, on Synthreo’s standard image - the tab shows the image name and its preinstalled packages. There is nothing to configure here; it exists so you know exactly where the agent’s tools run.
Limits
Section titled “Limits”Per-run guardrails - caps on steps, tokens, spend, and idle time:
| Field | Caps |
|---|---|
| Max steps / turn | How many actions the agent may take in one turn |
| Token budget | Total tokens a run may consume |
| Max spend (USD) | Hard cost ceiling per run |
| Idle timeout (hours) | How long a run may sit idle before ending |
Set these before enabling a Schedule trigger - they’re what keeps an unattended cadence inside a predictable budget.
Config: the canonical document
Section titled “Config: the canonical document”The Config tab holds the agent’s whole configuration as a single document: “The canonical config this agent saves. Editing the JSON updates every tab.”
- JSON is authoritative and editable - change it and select Apply, and every tab re-seeds from it.
- YAML is a read-only secondary view for easier reading.
This is also where advanced settings live that have no dedicated tab yet (for example the agent’s step/outcome structure).
Status and publishing
Section titled “Status and publishing”The header shows a Draft or Active badge with Save, Run, and Publish:
- Save stores your edits without changing what’s live.
- Run fires the agent now (“Run started”).
- Publish promotes the saved configuration to the live revision - triggers and other users always get the published revision.
Related
Section titled “Related”- Create your first Agent
- Schedule an Agent
- DAGs in Pylon - the classic node workflows