Discipline
Worldview
Discipline
The declared understanding of a domain — what entities exist, what authority structures govern them, what signals mean in context, and what constitutes legitimate action.
A worldview is authored by humans, versioned, and immutable once published. It is the intellectual substance that gives interpretation its meaning. Not a configuration. Not a model. A governed declaration of how a domain works.
The worldview is what makes a domain pack meaningful. The domain pack is how the worldview is expressed in executable form.
A worldview is authored by humans, versioned, and immutable once published. It is the intellectual substance that gives interpretation its meaning. Not a configuration. Not a model. A governed declaration of how a domain works.
The worldview is what makes a domain pack meaningful. The domain pack is how the worldview is expressed in executable form.
Appears in: Domain Studio · Worldview Engineering Session · Domain Packs · Interpretation Engine · INTERPRETIVE_WORLDVIEW.md
Worldview Engineering
Discipline
The discipline of encoding institutional knowledge as governed, versioned, auditable meaning infrastructure.
Worldview Engineers translate the implicit understanding that domain experts carry — what signals mean, what authority structures govern action, what constitutes legitimate behavior — into explicit worldviews that autonomous systems can operate within.
A new professional discipline that emerges wherever autonomous systems require governed interpretation before action. The role that makes structural governance possible at organizational scale.
Worldview Engineers translate the implicit understanding that domain experts carry — what signals mean, what authority structures govern action, what constitutes legitimate behavior — into explicit worldviews that autonomous systems can operate within.
A new professional discipline that emerges wherever autonomous systems require governed interpretation before action. The role that makes structural governance possible at organizational scale.
The expertise that previously lived in people's heads becomes governed infrastructure.
Appears in: Domain Studio · Positioning Series · What This Is Series · Emerging role narrative
Structural Governance
Discipline
Governance that is produced as a structural property of execution — not assembled in response to it. It exists before the action because the architecture requires it to. It is not a feature layered on top of a system. It is a precondition of the system operating at all.
Structural governance cannot be lost, reconstructed, or retroactively altered — because it was produced before the action, not because of it.
The three governance layers:
Access governance
— is this principal permitted to act?
Behavioral governance
— what did the agent do?
Structural governance
— what does this signal mean, before action?
Appears in: /structural-governance · Category positioning · Investor narrative
Authoring
Domain Studio
The authoring environment where worldviews are designed. Domain Studio guides domain experts through the Worldview Engineering Session — a structured four-phase process that produces a domain pack.
Every AI proposal and every human decision made during authoring is recorded in permanently separated fields. The H1-Authoring artifact is produced at session completion and written to the Human Decision Ledger.
Every AI proposal and every human decision made during authoring is recorded in permanently separated fields. The H1-Authoring artifact is produced at session completion and written to the Human Decision Ledger.
Domain Studio is where meaning is designed. Prism is where meaning is executed. No other system is permitted to decide what signals mean.
Appears in: A1 Architecture · domain_studio.md · H1-Authoring Addendum v0.1
Domain Pack
The technical artifact that encodes a worldview for a specific operational context. A domain pack declares signal types, event families, interpretation rules, and consequence mappings. Authored through the Worldview Engineering Session, versioned, and immutable once published.
Structure: four YAML files — naming_canon.yaml, normalized_event_families.yaml, product_capability.yaml, correlation_rules.yaml.
Structure: four YAML files — naming_canon.yaml, normalized_event_families.yaml, product_capability.yaml, correlation_rules.yaml.
Appears in: Domain Studio · Interpretation Engine · SKU Definition · Domain Pack Specification
Worldview Engineering Session
The structured four-phase authoring process in Domain Studio through which domain experts encode a worldview into a domain pack.
Phase 1— Vocabulary
Phase 2— Signal Classification
Phase 3— Capability Scope
Phase 4— Consequence Mapping
AI proposals and human decisions are permanently recorded in separately attributed fields — structurally separate and non-mergeable.
Appears in: Domain Studio · H1-Authoring Addendum v0.1 · Human Decision Ledger
H1-Authoring Artifact
The permanent, attribution-separated record of a Worldview Engineering Session. Contains every AI proposal and every human decision in permanently separated fields. Written to the Human Decision Ledger at session completion. Immutable.
The full provenance chain: H1-Authoring artifact → Domain Pack → Signal interpretation → Governed Meaning Artifact. Every link is ledgered and traceable.
The full provenance chain: H1-Authoring artifact → Domain Pack → Signal interpretation → Governed Meaning Artifact. Every link is ledgered and traceable.
Appears in: Human Decision Ledger · Ledger Authorization v0.1 · LA-I-8 invariant
H1-Decision Artifact
A human decision recorded as a signal with its own artifact class in the Human Decision Ledger. Operator authority decisions, delegations, standing instructions. Runtime-queryable by the interpretive loop for authority precedent evaluation.
The human decision is not just a record — it is a signal that informs future interpretation.
The human decision is not just a record — it is a signal that informs future interpretation.
Appears in: Human Decision Ledger · Ledger Authorization v0.1 · H1 Human Decision Surface
Core Runtime
Signal
Core
A machine-generated event that enters the S&P interpretation pipeline. A signal carries provenance — when it occurred, what system produced it, what it represents. Signals are not interpreted at intake. They are evidence. Meaning is produced by Prism — not by the signal itself.
Signals are universal. Domains interpret. The signal envelope structure is the same regardless of source — Azure Entra, GitHub, a drone sensor, a robot's proximity detector.
Signals are universal. Domains interpret. The signal envelope structure is the same regardless of source — Azure Entra, GitHub, a drone sensor, a robot's proximity detector.
Appears in: Signal intake surface · Signal adapter · Signal envelope · Universal Signal Layers v0.1
Prism
Core
The interpretation engine at the core of Signal & Prism. Every signal that crosses an action boundary passes through Prism before any execution path is reached.
Prism applies a declared worldview to an incoming signal and produces a governed meaning artifact. Deterministic — same signal, same worldview version, same artifact. Every time. No LLM in the interpretation path. No probabilistic inference.
Prism applies a declared worldview to an incoming signal and produces a governed meaning artifact. Deterministic — same signal, same worldview version, same artifact. Every time. No LLM in the interpretation path. No probabilistic inference.
No action exists without a prior interpreted meaning.
Appears in: Architecture · Zone 1 · Interpretive Loop · Runtime · Prism Specification
Governed Meaning Artifact
Core
The deterministic output of Prism's interpretation of a signal. Contains: interpretation finding, advisory state, boundary recommendation, authority posture, cited evidence, narrative explanation, domain pack version, worldview version, content hash.
Hash-verified at emission. Appended to the Artifact Ledger. Immutable. The governance record that exists before any action is taken — not reconstructed afterward.
Hash-verified at emission. Appended to the Artifact Ledger. Immutable. The governance record that exists before any action is taken — not reconstructed afterward.
Appears in: Zone 1 output · Artifact Ledger · interpretation_artifact_schema.py · Audit · Legal defensibility
Zone 1
Core
The fully deterministic interpretation pipeline. Signal intake through governed meaning artifact emission.
No LLM. No probabilistic inference. No generative content. Same signal, same domain pack version, same artifact. Every time.
Zone 1 is the governance guarantee — the record it produces is legally defensible because it is reproducible. Zone 1 never contacts external systems. Zone 1 output cannot be altered by Zone 2.
No LLM. No probabilistic inference. No generative content. Same signal, same domain pack version, same artifact. Every time.
Zone 1 is the governance guarantee — the record it produces is legally defensible because it is reproducible. Zone 1 never contacts external systems. Zone 1 output cannot be altered by Zone 2.
Appears in: Architecture · Determinism claim · Patent · Audit · A-Series Truths
Zone 2
Core
The generative advisory layer. Narrative generation, Atlas Copilot, AI Discovery in Domain Studio.
Zone 2 content is non-canonical — it does not affect the content hash and cannot corrupt the governed record. Zone 2 is explicitly separated from Zone 1 by architectural design. The two zones never merge.
Zone 2 is where humans explore meaning. Zone 1 is where meaning is produced.
Zone 2 content is non-canonical — it does not affect the content hash and cannot corrupt the governed record. Zone 2 is explicitly separated from Zone 1 by architectural design. The two zones never merge.
Zone 2 is where humans explore meaning. Zone 1 is where meaning is produced.
Appears in: Architecture · Copilot · Narrative layer · Atlas · A-Series Truths
Edge & Embodied
Micro-Prism
Edge
A constrained instantiation of the Prism interpretation engine designed for edge, embodied, or real-time deployment contexts where latency, determinism, or isolation are required.
Operates within a fixed, pre-declared autonomy envelope — a single domain, a pre-loaded worldview slice, bounded interpretation. Produces governed meaning artifacts locally. Synchronizes to the central Artifact Ledger when connectivity allows.
The same architecture as Prism. Deployed at the boundary of physical action.
Operates within a fixed, pre-declared autonomy envelope — a single domain, a pre-loaded worldview slice, bounded interpretation. Produces governed meaning artifacts locally. Synchronizes to the central Artifact Ledger when connectivity allows.
The same architecture as Prism. Deployed at the boundary of physical action.
Deployment contexts: Robots · Drones · Autonomous vehicles · Industrial automation · Security systems · Any environment where software causes physical effect
Appears in: Micro-Prism Autonomy Envelope v1.2 · Runtime Model v1.0 · Edge platform surface
Prism Field Unit
Edge
The physical form factor of the Micro-Prism — a small, ruggedized companion device attached to or co-located with an autonomous system.
Interprets sensor signals before any action is taken. Produces governed meaning artifacts locally. Synchronizes to the central Artifact Ledger when the system docks or regains connectivity. Operates offline. Deterministic by construction.
The governance record in physical form.
Interprets sensor signals before any action is taken. Produces governed meaning artifacts locally. Synchronizes to the central Artifact Ledger when the system docks or regains connectivity. Operates offline. Deterministic by construction.
The governance record in physical form.
Appears in: Edge platform surface · Autonomous environments · Hardware concept · Micro-Prism edge surface v0.1
Record / Composition
Artifact Ledger
Record
The immutable, append-only record of all governed meaning artifacts, human decisions, and agent registrations. Three co-equal ledgers — structurally parallel, connected by reference. No artifact ever deleted. Independent of any vendor. Cannot be reconstructed elsewhere.
The ledger is the organization's interpretive history — and its governance proof. Independent of any vendor. Cannot be reconstructed elsewhere.
The ledger is the organization's interpretive history — and its governance proof. Independent of any vendor. Cannot be reconstructed elsewhere.
Three co-equal ledgers + one introduced — structurally parallel, connected by reference:
Machine Artifact Ledger— governed meaning artifacts from signal interpretation
Human Decision Ledger— H1-Decision and H1-Authoring artifacts
Appears in: Downstream architecture · Switching cost · Audit lineage · human_decision_ledger.md
Prism Composer
Record
The mediator layer between the Artifact Ledger and Atlas. Composer reads from the ledger, traverses the artifact graph, assembles cross-domain projections, and pre-stages theater-ready views for Atlas to render.
Composer does not interpret signals. It assembles meaning that has already been governed. The projection is the input to the theater.
Composer does not interpret signals. It assembles meaning that has already been governed. The projection is the input to the theater.
Appears in: Downstream architecture · Canon Addendum: Downstream Architecture · Atlas rendering
Rendering
Atlas
Rendering
The theater rendering layer. Atlas receives pre-composed projections from Prism Composer and renders them as interactive theaters with interpretive optics.
Atlas does not generate meaning. It renders governed meaning. It is a theater, not an engine.
Atlas does not generate meaning. It renders governed meaning. It is a theater, not an engine.
Appears in: Rendering layer · Operator surface · live-atlas-optics · Zone 2 boundary
Theater
Rendering
A bounded rendering environment in Atlas where governed meaning artifacts are made visible and explorable. Each theater is scoped to a domain or scenario. The operator observes, explores, and evaluates — but does not modify the governed record.
Software Delivery
Knowledge System
Azure Identity— live
Appears in: Atlas · Demo · Operator experience · Theater rendering
Optic
Rendering
An interpretive lens applied to the theater. The same underlying governed meaning artifacts are rendered differently under each optic. The underlying reality does not change. The meaning shifts across optics.
Neutral— the event as it occurred
Identity— who was involved and their authority context
Code— what changed, beyond the surface label
Intent— declared versus actual
Authority— boundary and scope evaluation
Appears in: Atlas rendering · Theater presets · Operator navigation · live-atlas-optics
Legitimacy
Rendering
The interpretation finding that describes the relationship between an actor's action and the authority structures that govern it. Determined by Prism — not inferred by a model, not reconstructed from logs.
| Value | Color | Signal |
|---|---|---|
| would_allow | Slate | Within authority. Settled. |
| would_flag | Amber | Attention needed. Within boundary. |
| would_escalate | Orange | Elevated risk. Approaching boundary. |
| would_deny | Red | Outside authority. Boundary exceeded. |
Appears in: Interpretation engine · Governed meaning artifact · Atlas color model · Gravity stage
Gravity
Rendering
An Atlas theater preset that clusters nodes by legitimacy verdict. Actors with matching legitimacy pull toward each other. Outliers — would_deny, would_escalate — drift to the periphery.
The theater reads left to right from most concerning to most settled. Distance carries semantic weight. The operator's eye goes to the outliers first.
The theater reads left to right from most concerning to most settled. Distance carries semantic weight. The operator's eye goes to the outliers first.
Appears in: Atlas theater presets · CLAUDE.md Item 4b-i · Operator UX · Demo
Authority
Rendering
The explicit declaration of what a principal is permitted to do in a specific domain context. Authority structures are encoded in worldviews and enforced by Prism at interpretation time.
Authority is not inferred — it is declared. The authority posture of a governed meaning artifact describes whether the action falls within, at, or beyond the declared authority boundary.
Authority is not inferred — it is declared. The authority posture of a governed meaning artifact describes whether the action falls within, at, or beyond the declared authority boundary.
Appears in: Worldview · Domain pack · Interpretation engine · Atlas Authority optic · Ledger Authorization v0.1
Entry Briefing
Rendering
The summary generated at theater load that orients the operator before exploration begins. Produced deterministically from the Composer projection. Describes current theater state — entities present, attention areas, suggested exploration paths.
The entry briefing is Zone 2 — narrative summary, not a canonical artifact. It does not affect the governed record.
The entry briefing is Zone 2 — narrative summary, not a canonical artifact. It does not affect the governed record.
Appears in: Atlas · live-atlas-optics · Demo · Operator UX
Attention Areas
Rendering
The entities, sessions, or artifacts in the current theater that have legitimacy verdicts requiring operator focus. Surfaced in the Atlas right panel. Organized by legitimacy — would_deny and would_escalate appear first.
The operator's work begins with attention areas. The gravity stage makes them spatially visible — they cluster to the left periphery.
The operator's work begins with attention areas. The gravity stage makes them spatially visible — they cluster to the left periphery.
Appears in: Atlas · live-atlas-optics · Operator UX · Demo right panel
Customer Surfaces
Interpretive Control Platform
Customer Surface
The full product name for Signal & Prism in external communications. Encompasses all surfaces — Declare, Examine, Engage, Author, and the full Zone 1 interpretive pipeline — as a single governance infrastructure.
Use Interpretive Control Platform in press, partner, and investor materials. Interpretive control plane is the architectural description. Interpretive Control Platform is the product name.
Use Interpretive Control Platform in press, partner, and investor materials. Interpretive control plane is the architectural description. Interpretive Control Platform is the product name.
Established: May 16, 2026 · Appears in: External materials, press, briefing documents
Construction Declaration Record (CDR)
Customer Surface
The governed artifact produced by the Declare engine at declare.signalprism.com. Captures what an agent is authorized to do before it is deployed — the result of a four-phase conversational session covering six declaration dimensions: agent objective, signal surface, autonomous decisions, authority model, human gate, and success and drift.
Contains conversation lineage, gap inventory, and open items. The pre-deployment declaration record that travels forward through the authority chain. The building permit analogy: nobody gets a permit because they care about governance infrastructure. They get it because they need to build. The permit creates a permanent record, establishes what was authorized, and becomes the baseline for every subsequent inspection.
Contains conversation lineage, gap inventory, and open items. The pre-deployment declaration record that travels forward through the authority chain. The building permit analogy: nobody gets a permit because they care about governance infrastructure. They get it because they need to build. The permit creates a permanent record, establishes what was authorized, and becomes the baseline for every subsequent inspection.
The CLAUDE.md is a worldview declaration. The CDR governs it before the agent reads it.
Artifact prefix: cdr_{uuid} · Ledger: Agent Authority Ledger · Surface: declare.signalprism.com · Status: Live
Workflow Examination Record (WER)
Customer Surface
The governed artifact produced by the Examine engine at declare.signalprism.com. Captures the automation readiness of a workflow — the result of a five-phase examination covering automation readiness, integration readiness, and human gate.
Contains a step readiness map (GREEN/AMBER/RED per step), a first agent recommendation (specific, non-hedging), integration gaps, SLA gaps, notification gaps, and prerequisites. Tells the enterprise exactly where to start, what blocks deployment, and what the agent is authorized to handle.
Contains a step readiness map (GREEN/AMBER/RED per step), a first agent recommendation (specific, non-hedging), integration gaps, SLA gaps, notification gaps, and prerequisites. Tells the enterprise exactly where to start, what blocks deployment, and what the agent is authorized to handle.
Artifact prefix: wer_{uuid} · Surface: declare.signalprism.com · Status: Live
Provisional Authority Record (PAR)
Provisional · Concept
An auto-generated governed artifact produced when a signal arrives at S&P without a matching CDR or registration artifact in the Agent Authority Ledger. The PAR catches what wasn't declared.
When S&P observes a signal from an agent, machine, or human surface with no prior declaration, it generates a provisional toll booth entry. The PAR records what passed through, what it appeared to be authorized to do based on the signal, and opens a human follow-up workflow to confirm, correct, or escalate. The ledger permanently holds that something passed through — whether or not follow-up ever happens.
When S&P observes a signal from an agent, machine, or human surface with no prior declaration, it generates a provisional toll booth entry. The PAR records what passed through, what it appeared to be authorized to do based on the signal, and opens a human follow-up workflow to confirm, correct, or escalate. The ledger permanently holds that something passed through — whether or not follow-up ever happens.
S&P catches what wasn't declared and issues a provisional receipt — so the ledger always knows what passed through, whether or not anyone asked for governance.
provisional— generated, not yet acknowledged
notified— routed to named human or team
confirmed / corrected / escalated— human responded
resolved— superseded by full CDR or ara_
Artifact prefix: par_{uuid} · Ledger: AAL (provisional) · Status: Concept — PAR-OI-1 spec in progress
Forward Deployed Engineer (FDE)
Customer Surface
An engineer role emerging across Google, Anthropic, OpenAI, and Microsoft responsible for deploying AI capability into live enterprise infrastructure. FDEs build the connective tissue between AI capability and enterprise systems.
Every FDE is making undeclared worldview decisions at production speed — what the agent is authorized to do, what signals it should respond to, what authority model governs its actions. The Declare and Examine engines formalize those decisions into governed artifacts before deployment.
Every FDE is making undeclared worldview decisions at production speed — what the agent is authorized to do, what signals it should respond to, what authority model governs its actions. The Declare and Examine engines formalize those decisions into governed artifacts before deployment.
FDE connective tissue connects capability to infrastructure — not capability to authority. The governance gap is structural to the role.
Appears in: FDE professional practice, Declare engine, CDR, positioning series
Ledger
Machine Artifact Ledger (MAL)
Ledger
The first of three co-equal, append-only ledgers. Every governed meaning artifact produced by Zone 1 interpretation. Hash-verified. Lineage-traced to source signal and domain pack version. No artifact ever deleted. The MAL is not a section of the HDL or AAL — it is a separate, co-equal record.
Appears in: interpretation engine · artifact_ledger.py · Zone 1 pipeline
Human Decision Ledger (HDL)
Ledger
The second of three co-equal, append-only ledgers. Every human decision — authorization, escalation, acknowledgment, authoring. Two canonical artifact classes: H1-Decision and H1-Authoring.
LA-I-8 invariant: AI proposal and human decision blocks are permanently separated in all H1-Authoring artifacts. SESSION_STEPS carries AI proposals (immutable). Component state carries human decisions. Attribution separation is immutable at emission — never merged, collapsed, or re-attributed.
LA-I-8 invariant: AI proposal and human decision blocks are permanently separated in all H1-Authoring artifacts. SESSION_STEPS carries AI proposals (immutable). Component state carries human decisions. Attribution separation is immutable at emission — never merged, collapsed, or re-attributed.
Artifact classes: H1-Decision · H1-Authoring (ha_{uuid}) · Engagement artifacts (ea_{uuid})
Agent Authority Ledger (AAL)
Ledger
The third of three co-equal, append-only ledgers. Every agent registration — approved and denied. Delegation chain. Declaration tier. Supersession records when declarations change. The governed baseline that travels with an agent into every enterprise deployment.
Trust tier derived from declaration depth — never self-reported (ARS-I-3). S&P records. It does not certify. Also holds Provisional Authority Records (PAR) for ungoverned signals without a prior CDR.
Trust tier derived from declaration depth — never self-reported (ARS-I-3). S&P records. It does not certify. Also holds Provisional Authority Records (PAR) for ungoverned signals without a prior CDR.
Artifact prefixes: ara_{uuid} (approved) · ard_{uuid} (denied) · par_{uuid} (provisional)
Field Intelligence Ledger (FIL)
Ledger · Introduced
The fourth ledger alongside MAL, HDL, and AAL. Referential storage for FDE engagement context: session records, construction notes, engagement lineage, pattern references, and handoff records. Query pattern: by customer environment, agent class, and engagement stage.
The FIL closes the feedback loop that FDEs are hired to close — the context doesn't get handed off in a closeout doc, it gets declared into a permanent artifact.
The FIL closes the feedback loop that FDEs are hired to close — the context doesn't get handed off in a closeout doc, it gets declared into a permanent artifact.
Status: Introduced May 16, 2026 · Specification in progress (FIL-OI-1)
Pipeline
Interpretive Continuity
Core Runtime
How prior governed meaning artifacts inform current interpretation. The ledger is not a passive record — it is an active context source. When Prism interprets a new signal, the Ledger Query Contract surfaces a sealed prior context block containing relevant prior artifacts for the same entity or signal class.
Interpretation that ignores prior context is weaker than interpretation that can see the history of governed decisions. Interpretive continuity is the mechanism that makes the ledger compound in value over time.
Interpretation that ignores prior context is weaker than interpretation that can see the history of governed decisions. Interpretive continuity is the mechanism that makes the ledger compound in value over time.
Appears in: LQC · interpretation engine · interpretive_continuity.md · Status: Promoted from Deferred
Ledger Query Contract (LQC)
Core Runtime
The governed interface for reading the ledger within the interpretive loop. Defines four query types (Entity-Scoped Temporal, Artifact Lineage, Signal-Type-Scoped, Cross-Ledger) and six isolation guarantees.
The interpretive loop does not hold direct ledger access — it receives a sealed prior context block assembled by the LQC. This ensures interpretation remains deterministic and ledger access patterns cannot be manipulated at the point of interpretation.
The interpretive loop does not hold direct ledger access — it receives a sealed prior context block assembled by the LQC. This ensures interpretation remains deterministic and ledger access patterns cannot be manipulated at the point of interpretation.
Specification: ledger_query_contract.md v0.1 · Status: Promoted from Deferred