Pipeline model

A pipeline is an ordered list of blocks. During execution, blocks read from and write to an isolated variable store.

Variable scopes

Scope	Purpose
input.*	Data supplied to a pipeline run
data.*	Block output and response-scoped values
globals.*	Shared configuration values
bare name	User variables and captures, such as <TOKEN>

Use explicit variable names at integration boundaries. Avoid relying on legacy response aliases when a request block has a named response variable.

Status classification

Key Check evaluates explicit conditions and classifies a run as Success, Failure, Ban, Retry, or leaves it unclassified. Empty condition groups do not match, and numeric comparisons require valid numeric operands.