Skip to main content
Every Agentheya agent is defined by a set of configuration sections. This chapter documents each one field by field, in the same hierarchy as the canonical system schema. Each section page links back to the dashboard page where you edit it.
These pages reflect exactly what the configuration validator accepts — author against them precisely.

How configuration is stored

Each section is stored as JSON and validated on every write against its schema (additionalProperties:false, enums and required enforced, no coercion). Author against these pages exactly — unknown fields are rejected, not ignored.

Tier layering

Sections and individual fields can be gated by plan tier (x_min_tier). The schema is served layered to the caller’s plan: fields above your plan are stripped from reads and rejected on writes. A Tier column below and a badge on each field mark anything above the baseline. (Free trial = full Pro; Business unlocks the rest.)

All sections

Identity & purpose

Knowledge & memory

Tools & capabilities

Channels & interface

Commerce & policies

Operations

Nested sub-schemas

Two sections have rich sub-schemas documented on their own granular pages: