Skip to content

AAC Specification

Overview

The Agentic Automation Canvas (AAC) specification defines a structured data model for capturing comprehensive metadata about agentic automation projects. The specification ensures that all canvas data can be validated, transformed into standards-compliant RO-Crate packages, and integrated with FAIR Digital Object ecosystems.

Purpose

The AAC specification serves multiple purposes:

  1. Structured Data Capture: Provides a consistent schema for capturing project metadata, requirements, feasibility assessments, governance structures, and outcomes.

  2. Validation: Enables programmatic validation of canvas data to ensure completeness and correctness before export.

  3. Interoperability: Ensures generated RO-Crates follow established standards (RO-Crate 1.1, Schema.org, DCAT, PROV-O) for integration with research infrastructure.

  4. Documentation: Serves as a reference for understanding the data model, field semantics, and relationships between entities.

  5. Tooling: Enables development of CLI tools, API integrations, and other interfaces that work with AAC data.

Scope

The specification covers:

  • Project Definition: Core project metadata, objectives, stages, and domain information
  • User Expectations: Requirements, stakeholders, and benefit metrics
  • Developer Feasibility: Technical risk assessment, TRL levels, model selection, and implementation considerations
  • Governance: Governance stages, agents, milestones, and compliance standards
  • Data Access: Dataset metadata, access rights, sensitivity levels, and DUO terms
  • Outcomes: Deliverables, publications, and evaluation results

Specification Structure

The specification is organized into several sections:

  • Core Concepts: Fundamental concepts and entities (canvas, task, contract, evaluation, provenance)
  • Conformance: Conformance requirements, versioning policy, and deprecation procedures
  • Schema Reference: Detailed field-by-field reference documentation

Schema Version

The current schema version is 0.10.1 (Beta). The schema follows semantic versioning:

  • Major versions (1.0, 2.0): Breaking changes, incompatible with previous versions
  • Minor versions (0.10, 0.11): Pre-release versions, may include breaking changes
  • Patch versions (0.10.1, 0.10.2): Bug fixes, documentation updates

Beta Testing Phase

Version 0.10.1 is a beta release for testing and feedback. The schema structure and fields may change before the stable 1.0.0 release. We welcome feedback and contributions during this phase.

See the Conformance section for details on versioning and compatibility.

Validation

All AAC canvas data MUST conform to the JSON Schema defined at:

https://w3id.org/aac/schema/aac.schema.json

Validation can be performed:

  • Programmatically using JSON Schema validators
  • Via the CLI validator
  • Through the web application's built-in validation

The AAC specification builds on and integrates with:

  • RO-Crate 1.1: Research Object Crate specification for packaging research outputs
  • Schema.org: Vocabulary for structured data on the web
  • W3C DCAT: Data Catalog Vocabulary for describing datasets
  • W3C PROV-O: Provenance Ontology for describing activities and agents
  • P-Plan: Plan Ontology for describing plans and steps
  • FRAPO: Funding, Research Administration & Projects Ontology
  • DUO: Data Use Ontology for data use restrictions

See the schema mappings for detailed mapping documentation.