What AWO defines
AWO is the operational method that turns ARI governance and the Neurotransparency Specification into concrete workflows. It defines which roles exist, which artifacts they must produce, and how a research run progresses from idea to attested result.
Think of AWO as the “how” layer: if a claim is made under Waveframe Labs, AWO specifies the path that claim must have taken.
Core Artifacts
Minimum required artifacts
- Run manifest: machine-readable description of the run, roles, models, and parameters.
- Attestation log: human approvals, notes, and override justifications.
- Evidence bundle: outputs, prompts, and intermediate reasoning artifacts.
- Metadata headers: standardized headers applied to all relevant files.
- Governance log: record of exceptions, failed checks, and corrective actions.
Role Separation
AWO enforces separation between at least three conceptual roles:
- Builder: assembles workflows, prompts, and experiments.
- Reviewer: audits artifacts, checks claims, and approves or rejects runs.
- Orchestrator: coordinates automation and connects AWO to CRI-CORE.
These roles may be implemented by one or more humans plus AI systems, but their responsibilities must remain distinguishable in the logs.
Alignment with Doctrine & ARI
- AWO encodes the Neurotransparency Specification as explicit workflow rules.
- AWO complies with ARI’s governance decisions around approvals, oversight, and exceptions.
- CRI-CORE is expected to treat AWO manifests as contractual input for enforcement.
Where to Read the Method Specification
- GitHub repository: Waveframe-Labs/Aurora-Workflow-Orchestration
- Specification PDF: see the latest release artifacts in the repository.