author
Normative rules for identifying the human or organizational author behind a piece of content, including how author IDs and references are conveyed to AI agents.
MSP-1 Specification
MSP-1 Core Specification
This section defines the normative behavior of the MSP-1 — Mark Semantic Protocol — so that humans, AI agents, and tooling can interpret MSP-1 metadata consistently across any compliant site.
Identity, authority & trust
Terms that express who is speaking, on whose behalf, and how trust and provenance are represented at the metadata layer.
authority
Defines how MSP-1 expresses authoritative sources, including site-level and document-level authority relationships and how agents should interpret them.
provenance
Specifies how to declare origin, lineage, and upstream sources for content, enabling AI agents to trace where information comes from.
reviewer
Describes how to indicate human or organizational reviewers, including review roles, scopes, and what “reviewed” means under MSP-1.
trust
Outlines the trust-related signals exposed by MSP-1, including levels of verification and how AI agents should weigh them.
role
Provides a standardized way to describe roles (author, editor, reviewer, owner, etc.) associated with content and entities within an MSP-1 implementation.
site
Defines the concept of a “site” within MSP-1, including identifiers, scope boundaries, and how multi-site ecosystems are expressed.
Document lifecycle, versioning & compliance
Terms that govern how content evolves over time, how versions are expressed, and how MSP-1 compliance is communicated.
canonical
Normative guidance for declaring a canonical representation of content, including how agents should resolve duplicates or variants.
discovery
Defines how MSP-1 declarations are discovered, anchored to the canonical
/.well-known/msp.json endpoint, and how agents should avoid
inference-based filename guessing.
compliance
Defines MSP-1 compliance levels, required vs. optional features, and how sites should advertise their conformance state.
revision
Describes how to represent discrete revisions of a resource, including revision identifiers and their relationship to versions.
revisionDate
Specifies the expected semantics and formatting of revision timestamps, enabling consistent temporal reasoning.
revisionNotes
Defines how MSP-1 captures human-readable explanations of what changed between revisions for both humans and agents.
revisionVersion
Explains how to map revisions onto logical version numbers, including semantic versioning conventions within MSP-1.
version
Provides a normative model for version identifiers at the protocol, site, and resource levels, including MSP-1.0.x semantics.
Content structure, intent & addressing
Terms that describe what a resource is, how it is structured, and how AI agents should interpret its intent and placement.
description
Defines how MSP-1 expects concise, high-signal descriptions to be authored for AI-friendly comprehension and summarization.
id
Normalizes identifiers used within MSP-1, including stable IDs for sites, pages, sections, and conceptual entities.
intent
Specifies how to articulate the underlying intent or purpose of a resource so AI agents can align answers with author goals.
interpretiveFrame
Defines interpretive context for content — constraints, assumptions, or lenses that should shape how agents read a resource.
name
Describes how human-readable names and titles should be expressed for maximum clarity and disambiguation.
page
Defines what constitutes a “page” in MSP-1 terms, including routing, scoping, and how pages relate to sections and sites.
parent
Specifies parent–child relationships between resources, enabling nested hierarchies and structural reasoning.
protocol
Clarifies how MSP-1 as a protocol is declared, including protocol identifiers, supported ranges, and negotiation.
section
Describes intra-page structural units and how sections are exposed to AI agents for finer-grained understanding.
type
Explains how MSP-1 uses type declarations to categorize resources, content blocks, and entities in a consistent way.
url
Defines URL expectations within MSP-1, including stability, cleanliness, and how agents should resolve relative vs. absolute URLs.