Atlas Lexicon

This lexicon defines the core terms used across Project Atlas. Link to these anchors when describing systems, patterns, and playbooks.

Canonical

The authoritative representation of a dataset, configuration, or interface. Canonical forms are the reference point for diffing, validation, and downstream automation.

Diff/apply

A workflow that computes differences between the current state and the desired state, then applies changes in a controlled, auditable way. Diff/apply workflows make automation repeatable and safe.

Idempotent compliance pass

A repeatable automation run that reasserts a baseline without causing harm when executed multiple times. The goal is to keep systems aligned with policy through repeated checks and fixes.

Audit artifacts

Evidence produced by a system run: diff reports, logs, exported CSVs, change summaries, and validation outputs. Audit artifacts make outcomes verifiable and reduce guesswork during incident response.

Mapping layer

A translation layer that maps source system labels to target system IDs, codes, or normalized values. Mapping layers keep integrations stable when source values are messy or inconsistent.

Reality → Action trace

A shared view of how a system turns real-world inputs into reliable actions, from ingestion through storage, automation, interfaces, and adoption. The trace makes system boundaries and handoffs explicit.

Operational adoption loop

The cycle that makes systems stick: playbooks, ownership, training, feedback, and continuous improvement. Adoption is a loop, not a one-time rollout.