Point
People and agents share one semantic context
The portable Atlas should give people and software agents the same identities and provenance while keeping operational authority separate.
- Posture
- asserted
- Lifecycle
- active
A person and a software agent should encounter the same Point identity, primary Map, context records, state, relations, Content, References, and source provenance. Separate machine-only and human-only meanings would weaken portability and make project decisions harder to inspect.
Shared context does not imply shared capability. A reader can learn that a deployment is intended or that a Check applies without receiving credentials, file access, network access, or permission to execute the action. Applications must combine Atlas interpretation with authority supplied independently by their caller and environment.
Shared navigationCommon Point identities and source provenance are the shared navigation contract for both human and software readers.
Project contextA common semantic model lets both audiences inspect the same connected decisions and evidence.
Human and agent authoring should preserve the same identities, contribution boundaries, evidence roles, and review outcomes while giving multiple writers one stable subject.
A software agent must not create a parallel machine-only interpretation of the Atlas. Its proposed changes use the same source records and Checks that a person reviews, while any file or network capability remains explicit in the caller rather than encoded as semantic context.
Point identifiers also give several agents a stable coordination unit. Agents can name the same subject, inspect the same resolved records, contribute different Map-local consequences, and submit changes for review without exchanging free-form summaries that quietly redefine the claim.
Contribution boundariesHuman and agent writers, including parallel agents, must use the same Point identity and anchor-versus-context boundary so their changes assemble together.
AuthorityA shared authoring model does not give either kind of writer additional operational permission.
Shared semantics let different tools and users exchange context, but ecosystem implementations still need transparent provenance and separate capability controls.
Independent applications should not need a proprietary agent representation beside the human-readable Atlas. They can optimize interfaces and workflows, but exchanged data must preserve the same Point identity and contribution sources so people can audit what an agent interpreted or changed.
PortabilityShared identities provide an interchange vocabulary across implementations and users.
IntegrationTransparent provenance and capability controls are required when tools present that shared context.