Orange Canonical Precedents
Canonical precedents distilled from Orange, capturing reusable governance rules that apply across consequential AI agent actions.
Distilled from Orange judgments
Orange canonical precedents are derived from recurring authority, evidence, release, and execution-boundary failures observed within the Orange system.
Written as reusable rules
Each canonical precedent captures a general governance rule that can guide future cases across different tools, systems, and environments.
Expanded through GAIR volumes
GAIR stands for Global Agent Incident Registry. Its volumes organize incidents, engineering failures, enterprise risks, and positive controls under these foundational precedents.
GAIR Case Volumes
GAIR Volume 01
Public High-Impact Incidents
GAIR-001 – GAIR-010
GAIR Volume 02
Open-Source Agent Engineering Failures
GAIR-011 – GAIR-020
GAIR Volume 03
Enterprise Risk and Compliance Incidents
GAIR-021 – GAIR-030
GAIR Volume 10
Sovereignty Restoration and Positive Control Patterns
The first nine GAIR volumes document how AI agent actions fail across authority, evidence, execution boundaries, collaboration, audit systems, and time. Volume 10 turns those failures into reusable positive controls. It defines patterns for deny-by-default execution, evidence-first closure, human override, external interruption, strict release paths, and restored human sovereignty.