Agent Configuration#

Agent configuration defines the GenAI agents that power the Chat dashboard in Squirro Neo: the models behind them, the capabilities they are allowed to use, the knowledge they can reach, and the preferences that govern their behavior.

Overview#

Explains how agent configuration works: what you author, how settings are applied, and how the platform picks a value when multiple sources disagree.

Learn more

Prerequisites#

Everything to have in place before you start: a target project, the genai command line tool, a project Administrator refresh token, and registered credential handles.

Learn more

Installation#

How to install the genai command line tool, including its requirements and how to confirm it is working.

Learn more

Quick Start#

A step-by-step guide to writing, validating, and deploying your first configuration, then checking that it works as expected.

Learn more

Authoring Agents#

Covers what you can configure on an agent, such as its instructions and default settings, and the difference between agents a user can pick and agents that work behind the scenes.

Learn more

Scheduled Tasks and Approvals#

How to schedule an agent to run on a recurring schedule, and how to gate the tools it may call behind human approval.

Learn more

Resource Reference#

The complete field reference for every configuration kind: required and optional fields, types, allowed values, and a minimal example for each. Also covers the constraint predicate syntax and how credential handles are registered.

Learn more

Preference Reference#

The complete catalog of preference paths across the inference, capability, source, and trait namespaces, with the type, allowed values, and default for each.

Learn more

CLI Reference#

A complete reference for the genai command line tool: create, check, deploy, and inspect a configuration.

Learn more

API Reference#

The HTTP API behind the genai command line tool, for building your own tools or automating deployment.

Learn more

Troubleshooting#

Common failures when authoring a configuration, deploying it, calling the API, and running scheduled tasks, each with its symptom, cause, and fix.

Learn more