SHAREPLANE AGENT CONTEXT Schema: shareplane-context/2.0 Schema URL: https://next.shareplane.malott.ai/schemas/agent-package.schema.json Projection: Generated public-safe plain text. Not canonical Markdown. Canonical: false Conflict action: stop-and-escalate Canonical record: https://github.com/pinklon/pinklon-shareplane-next/tree/33d227f6000da2491f19209edde916d8846f287f/content/artifacts/architecture-and-intent/artifact.json Canonical record SHA-256: faa4745d8551606d2c182672f876730cae4b036c4325d8ccf90ce10de314c402 Source content SHA-256: 43177a3463ba7f57194abb92792f3774cbd4d8d1c11c5788875ed64074024300 Generation receipt: https://next.shareplane.malott.ai/build-receipt.json Content role: artifact-content Content trust: untrusted-data Instructions allowed: false Operational authority: none IDENTITY Artifact ID: artifact:architecture-and-intent Slug: architecture-and-intent Canonical URL: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/ Title: Architecture and Intent Abstract: Your architecture diagram shows what the system can reach. It usually does not show what the system should trust, who has authority, how contradictions are resolved, what may be changed, or how far the consequences can spread. Author: Tony Malott Author URL: https://malott.ai Published: 2026-07-17 Updated: 2026-07-17 Format: architecture-thesis Privacy: public-safe Topics: - enterprise-architecture - ai-governance - source-authority - human-agent-collaboration - provenance - systems-thinking Audience: - architects-and-systems-thinkers - review-boards-and-implementation-teams - executives-and-architecture-leaders PROVENANCE Posture: owner-approved-public-safe-presentation-family-v02 Private sources used: true Private sources published: false Public-safe boundary: This package uses generalized examples involving architecture records, policies, controls, services, operational assets, repositories, databases, evidence, incidents, approvals, lifecycle, and organizational knowledge. It contains no employer or customer names, private estate counts, regulated records, internal control identifiers, private topology, internal URLs, named private owners, credentials, secrets, or security-sensitive production configuration. The artifact expresses Tony Malott's architecture doctrine and a proposed review method. It does not claim enterprise approval, regulatory certification, or universal applicability. CLAIMS Claim: claim:architecture-and-intent:c01 Posture: Tony doctrine and field observation Text: A component diagram can be technically correct while omitting truth hierarchy, authority, conflict handling, mutation boundaries, and stop conditions. Support: - source:architecture-and-intent:s01 - source:architecture-and-intent:s03 Claim: claim:architecture-and-intent:c02 Posture: Tony doctrine supported by security and provenance practice Text: Retrieval relevance does not establish source authority or approval. Support: - source:architecture-and-intent:s02 - source:architecture-and-intent:s05 Claim: claim:architecture-and-intent:c03 Posture: Tony doctrine Text: Recent information can remain wrong, unapproved, out of scope, or nonauthoritative. Support: - source:architecture-and-intent:s05 Claim: claim:architecture-and-intent:c04 Posture: Tony doctrine supported by least-privilege practice Text: Capability is not authority. Tool access does not itself create an organizational mandate. Support: - source:architecture-and-intent:s03 Claim: claim:architecture-and-intent:c05 Posture: Current security guidance Text: Retrieved or external content can carry instructions that alter model behavior, and RAG alone does not remove prompt-injection risk. Support: - source:architecture-and-intent:s02 Claim: claim:architecture-and-intent:c06 Posture: Current security guidance Text: Agentic risk should be constrained through narrow tools, least privilege, downstream authorization, approval for high-impact actions, monitoring, and rate limiting. Support: - source:architecture-and-intent:s03 Claim: claim:architecture-and-intent:c07 Posture: Current legal requirement in applicable EU AI Act contexts Text: Human oversight should be proportional to risk and support interpretation, override, reversal, interruption, and safe stopping. Support: - source:architecture-and-intent:s04 Claim: claim:architecture-and-intent:c08 Posture: Foundational standard Text: Provenance can represent derivation and processing history and support assessment of quality, reliability, and trustworthiness. Support: - source:architecture-and-intent:s05 Claim: claim:architecture-and-intent:c09 Posture: Tony doctrine and architectural inference Text: In context-as-code environments, changes to trusted context can shape future agent behavior. Support: - source:architecture-and-intent:s01 - source:architecture-and-intent:s05 Claim: claim:architecture-and-intent:c10 Posture: Tony doctrine supported by risk and security guidance Text: Consequential agent capabilities require explicit scope, reversibility, validation, observability, rollback, evidence, and stop conditions. Support: - source:architecture-and-intent:s01 - source:architecture-and-intent:s03 - source:architecture-and-intent:s04 Claim: claim:architecture-and-intent:c11 Posture: Tony doctrine supported by human-oversight requirements Text: Human judgment should be placed at explicit authority transitions rather than represented as a decorative review box. Support: - source:architecture-and-intent:s04 Claim: claim:architecture-and-intent:c12 Posture: Tony doctrine and proposed review method Text: A consequential AI architecture review should show both technology architecture and architecture of intent, then bind them through an alignment matrix. Support: - None declared. PUBLIC SOURCES Source: source:architecture-and-intent:s01 Title: NIST AI 600-1: Generative Artificial Intelligence Profile Type: public-source-dossier-entry Role: Current authoritative risk-management context for governance, provenance, testing, information integrity, information security, and human-AI configuration. Locator: https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.600-1.pdf Source: source:architecture-and-intent:s02 Title: OWASP LLM01: Prompt Injection Type: public-source-dossier-entry Role: Current security guidance for direct and indirect prompt injection, including retrieved external content and the limits of RAG as a mitigation. Locator: https://genai.owasp.org/llmrisk/llm01-prompt-injection/ Source: source:architecture-and-intent:s03 Title: OWASP LLM06: Excessive Agency Type: public-source-dossier-entry Role: Current security guidance for narrow tools, least privilege, downstream authorization, human approval for high-impact actions, monitoring, and rate limiting. Locator: https://genai.owasp.org/llmrisk/llm062025-excessive-agency/ Source: source:architecture-and-intent:s04 Title: Regulation (EU) 2024/1689, Artificial Intelligence Act Type: public-source-dossier-entry Role: Current legal authority for risk-proportionate human oversight, including interpretation, override, reversal, interruption, and safe stopping. Locator: https://eur-lex.europa.eu/eli/reg/2024/1689/oj Source: source:architecture-and-intent:s05 Title: W3C PROV Overview Type: public-source-dossier-entry Role: Foundational provenance authority retained as an explicit evidence-floor exception for entities, activities, derivation, versioning, reproducibility, and trust assessment. Locator: https://www.w3.org/TR/prov-overview/ RELATIONSHIPS Relationship: relatedTo Target: artifact:architecture-review-plane-interactive-concepts Posture: owner-approved-hosted-uat-relationship Evidence: Tony owner UAT approval in the Issue #109 execution session, 2026-07-17 PRESENTATION FAMILY Family ID: presentation-family:architecture-and-intent Semantic authority SHA-256: b524810904c2842bd52cbcb95cf1a35fbb6fe38ef2bf0c746261427dee6483a4 Render mode: shareplane-shell-exact-payloads-v1 All presentations share one semantic authority and differ only in visual and reading grammar. Chooser route: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/presentation-family/ Variants: - The Two-Plane Map (presentation-variant:architecture-and-intent:two-plane-map) Reader job: Architects and systems thinkers Visual grammar: architecture map (visual-grammar:spatial-architecture-map) Route: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/presentations/two-plane-map/ SHA-256: 883b454cf847664ac905309c7c044efda78e7bec831ca283a56f61ddceb5384b Bytes: 39942 - The Architecture Review Dossier (presentation-variant:architecture-and-intent:architecture-review-dossier) Reader job: Review boards and implementation teams Visual grammar: governed review dossier (visual-grammar:architecture-review-dossier) Route: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/presentations/architecture-review-dossier/ SHA-256: b3136c393bc06f11d3fdb189d6c136203b17c8228e972307b4d5222b9060bd09 Bytes: 34657 - The Missing Diagram (presentation-variant:architecture-and-intent:missing-diagram) Reader job: Executives and architecture leaders Visual grammar: editorial argument (visual-grammar:restrained-editorial-essay) Route: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/presentations/missing-diagram/ SHA-256: 42a6cedd0c7d5343ba0c0b7b2cc67e8dc3ed06954752fb44a15d6cd8805b6684 Bytes: 34963 Preference posture: browser-local persistence is permitted; server telemetry is prohibited; preference has no semantic authority. ARTIFACT CONTENT [HEADING 1] Architectureand Intent [PARAGRAPH] By Tony Malott [PARAGRAPH] The system diagram is only half the architecture. Most AI diagrams explain the stack. These three views explain the missing plane: what the system means, trusts, may decide, may change, and must refuse. [HEADING 2] Three ways into one thesis [PARAGRAPH] No winner is hidden in the interface. Pick the presentation that fits the job. All three preserve the same semantic authority. [PARAGRAPH] Architects and systems thinkers [HEADING 2] The Two-Plane Map [PARAGRAPH] Start with the system model. See how intent governs the technical stack through explicit bindings. [PARAGRAPH] Review boards and implementation teams [HEADING 2] The Architecture Review Dossier [PARAGRAPH] Use the contracts, conflict path, action ladder, and alignment matrix as an applied review method. [PARAGRAPH] Executives and architecture leaders [HEADING 2] The Missing Diagram [PARAGRAPH] Read the argument as a focused executive essay with visual models introduced only where they earn their place. [BLOCKQUOTE] Capability is not authority. [PARAGRAPH] The technical stack determines what the system can do. The architecture of intent determines what it should do, what it may do, and when it must stop. [PARAGRAPH] Bring me both diagrams. PUBLIC SURFACES Human page: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/ Metadata JSON: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/artifact.json Receipt: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/receipt.json Context: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/context.txt Agent-package manifest: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/agent-package.json Agent-package ZIP: https://next.shareplane.malott.ai/artifacts/architecture-and-intent/agent-package.zip Collection catalog: https://next.shareplane.malott.ai/catalog.json Graph: https://next.shareplane.malott.ai/graph.json Agent index: https://next.shareplane.malott.ai/llms.txt