@summit-cognitive/scsfs

Summit Cognitive Standard Field Set — The foundational schema for the Cognitive Admissibility Layer.

scsfs defines the canonical entity, relationship, and provenance schemas required to transform raw intelligence into admissible decisions. It is the formal specification for the Cognitive Admissibility Certificate (CAC).

The Admissibility Boundary

While existing tools (MF0-1984, Autoresearch, StegoForge) generate intelligence outputs, they often fail the standards required for high-stakes decision-making. scsfs provides the structural enforcement needed for:

• Determinism: Ensuring reasoning paths do not drift over time.
• Traceability: Cryptographic lineage from origin to inference.
• Defensibility: Explicitly tracking competing hypotheses and dissent.
• Replayability: The ability to reconstruct the exact cognitive state at the moment of decision.

Install

npm install @summit-cognitive/scsfs

Key Components

| Module | Description | |--------|-------------| | core/cac | Cognitive Admissibility Certificate — The flagship proof artifact for decision integrity. | | gate/validator | Admissibility Gate — Enforcement logic that evaluates entities against provenance and policy. | | adapters/* | Reference Adapters — Ingestion contracts for MF0-1984, Autoresearch, and StegoForge. | | core/base | CanonicalEntityBase, CanonicalRelationship, PolicyLabels, ProvenanceChain, RelationshipStrength, AnalystAnnotation, enums | | entities/person | PersonEntity, PersonEntityHelpers | | entities/organization | OrganizationEntity, OrganizationEntityHelpers | | entities/physical/location | LocationEntity, LocationEntityHelpers | | entities/physical/asset | AssetEntity, AssetEntityHelpers | | entities/physical/vehicle | VehicleEntity, VehicleEntityHelpers | | entities/physical/device | DeviceEntity, DeviceEntityHelpers | | entities/physical/infrastructure | InfrastructureEntity, InfrastructureEntityHelpers | | entities/digital/account | AccountEntity, AccountEntityHelpers | | entities/digital/communication | CommunicationEntity, CommunicationEntityHelpers | | entities/digital/document | DocumentEntity, DocumentEntityHelpers | | entities/digital/sensor | SensorEntity, SensorEntityHelpers | | entities/financial/financial-instrument | FinancialInstrumentEntity, FinancialInstrumentEntityHelpers | | entities/intelligence/indicator | IndicatorEntity, IndicatorEntityHelpers | | entities/intelligence/threat | ThreatEntity, ThreatEntityHelpers | | entities/intelligence/ttp | TTPEntity, TTPEntityHelpers | | entities/intelligence/incident | IncidentEntity, IncidentEntityHelpers | | entities/intelligence/vulnerability | VulnerabilityEntity, VulnerabilityEntityHelpers | | entities/intelligence/campaign | CampaignEntity, CampaignEntityHelpers | | entities/intelligence/narrative | NarrativeEntity, NarrativeEntityHelpers | | entities/operational/event | EventEntity, EventEntityHelpers | | entities/operational/case | CaseEntity, CaseEntityHelpers | | entities/operational/runbook | RunbookEntity, RunbookEntityHelpers | | entities/governance/license | LicenseEntity, LicenseEntityHelpers | | entities/governance/authority | AuthorityEntity, AuthorityEntityHelpers | | entities/epistemology/claim | ClaimEntity, ClaimEntityHelpers | | entities/epistemology/evidence | EvidenceEntity, EvidenceEntityHelpers | | entities/epistemology/hypothesis | HypothesisEntity, HypothesisEntityHelpers | | entities/phenomenology/episode | EpisodeEntity, EpisodeEntityHelpers | | entities/phenomenology/perspective | PerspectiveEntity, PerspectiveEntityHelpers | | entities/phenomenology/mediation | MediationEntity, MediationEntityHelpers | | er/types | ERDecision, ResolutionCluster, ERMatchScore, ERThresholds, ERPrecisionMetrics |

Sandbox Demo

You can run the end-to-end admissibility demonstration to see how the Gate enforces trust boundaries:

npm run demo

The demo covers high-trust certification, policy-based rejection (missing legal basis), and confidence-floor enforcement for classified data.

The Cognitive Admissibility Certificate (CAC)

The CAC is the primary output of a Summit-compliant pipeline. It wraps every entity and relationship in a verifiable envelope that proves:

1. Hash-at-origin integrity.
2. Policy-compiled enforcement (Clearance, Sensitivity, Retention).
3. Reasoning-chain linkage (No "black box" inferences).

Usage

import {
  CognitiveAdmissibilityCertificate,
  CanonicalEntityBase,
  ProvenanceChain,
  SensitivityLevel,
  VerificationStatus,
} from '@summit-cognitive/scsfs';

Policy Enforcement

Every artifact in scsfs carries mandatory PolicyLabels to ensure regulatory and operational compliance is baked into the data model:

interface PolicyLabels {
  origin: string;
  sensitivity: SensitivityLevel;    // PUBLIC | INTERNAL | CONFIDENTIAL | RESTRICTED | TOP_SECRET
  legalBasis: string;
  needToKnow: string[];
  retentionClass: RetentionClass;   // TRANSIENT | SHORT_TERM | MEDIUM_TERM | LONG_TERM | PERMANENT
}

Build & test

npm run build       # compile TypeScript → dist/
npm run typecheck   # type-check without emitting
npm test            # run vitest test suite
npm run test:watch  # watch mode

License

MIT © Summit Cognitive, Inc.