Ontology packs

Ontology packs extend the MIF entity vocabulary for a specific domain. Each pack supplies a *.ontology.yaml that declares namespaces, entity types, relationships, and discovery patterns. Binding an ontology pack lets the research engine recognize, classify, and relate domain entities found in sources.

Ontology packs have no external dependencies and no SKILL.md — they are data packs, not skill packs.

The vocabulary is layered in three tiers: the domain-neutral generic core (mif-base / mif-generic / shared-traits, the MIF-compliant always-on layer) → engineering-base (shared engineering supertypes — MIF-compliant, opt-in via extends, never bound directly) → the bindable domain packs below.

For control-plane mechanics see Packs and Plugins.

Enabling and binding an ontology

Ontology packs use a different control surface from the skill/channel/genre plugin packs. They are declared in harness.config.json ontologies[] (an id plus an enabled flag), not in packs[], so scripts/pack-toggle.sh does not apply to them. Enabling an ontology is two steps:

Set its enabled flag to true in ontologies[] (the per-pack “Enable” command below does exactly this for that pack’s id).
Bind the enabled ontology to a research topic with the /ontology-review command, whose deterministic engine is scripts/ontology-review.sh. Binding is what lets findings in that topic resolve to the ontology’s entity types; per-finding classification is handled by scripts/resolve-ontology.sh.

resolve-ontology.sh requires yq, jq, and ajv (see dependencies).

biology-research-lab

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/biology-research-lab/

Purpose

Provides an entity vocabulary covering the full lifecycle of an academic biology research lab: personnel, grants, experiments, samples, data, publications, and compliance (IRB / IACUC / IBC). Sources include NIH, NSF, OHRP, OLAW, FAIR Data Principles, and the CRediT Contributor Roles Taxonomy.

Domain

Academic biology research labs and their operational, funding, and compliance contexts.

Entities, relationships, and traits

Entity types: principal-investigator, lab-member (postdoc / graduate-student / technician / lab-manager / other), collaborator, grant, grant-submission, grant-report, project, protocol (cell-culture / molecular-biology / imaging / animal / computational / other), experiment, sample (cell-line / tissue / dna / rna / protein / plasmid / organism / other), reagent, equipment (microscope / centrifuge / sequencer / flow-cytometer / other), dataset, publication, manuscript-submission, irb-protocol, iacuc-protocol, ibc-protocol, training-record.

Key relationships: leads, funded_by, uses_protocol, produces, covered_by, collaborates_with, cites_grant, uses_sample, uses_equipment.

Traits applied: lifecycle, contactable, certified, renewable, auditable, inventoried, maintainable, versioned, owned, reviewed, scheduled, measured, budgeted.

Discovery patterns: recognizes NIH grant mechanism codes (R01, R21, K99, F31, T32, U01), ORCID identifiers, experimental keywords (PCR, qPCR, Western, assay), compliance keywords (IRB, IACUC, IBC, BSL), and publication identifiers (DOI, PMID).

When to bind

Bind biology-research-lab when researching academic research lab operations, life sciences grant landscapes, laboratory compliance requirements, or research data management.

Enable

jq '(.ontologies[] | select(.id=="biology-research-lab") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — topics must explicitly enable and bind; never auto-applied to non-biology-lab topics.
Extends mif-base v1.0.0 and shared-traits v1.0.0; binding is fail-closed — resolve-ontology.sh and validate-concordance.sh abort the entire corpus if either extends target is missing or mistyped.
Scoped to academic biology research labs; entity types do not apply to engineering, legal, or market-research topics.
Compliance sub-types (IRB / IACUC / IBC) are domain-specific and resolve only within bound biology-lab topics.

Goals

Supplies entity vocabulary covering the full biology-lab lifecycle: personnel (PI, postdoc, graduate-student, technician, lab-manager), grants, experiments, samples, reagents, equipment, publications, and compliance protocols.
Enables recognition of NIH grant mechanism codes (R01, R21, K99, F31, T32, U01), ORCID identifiers, assay keywords, compliance identifiers (IRB, IACUC, IBC, BSL), and publication identifiers (DOI, PMID) in research sources.
Typed findings validate fail-closed against the MIF schema on binding, providing provenance and completeness guarantees.
Supports compliance lifecycle tracking and research data management (FAIR principles, CRediT contributor roles) within bound topics.

data-engineering

Version: 0.2.0 | Kind: ontology

Source: packs/ontologies/data-engineering/

Purpose

Provides an entity vocabulary for the data engineering domain: data contracts, data products, governance policies, data quality, storage architectures, and pipeline patterns. It extends engineering-base, inheriting the shared engineering supertypes and keeping only its data-specific types.

Domain

Data engineering teams, data platform engineering, and modern data infrastructure.

Entities, relationships, and traits

Namespaces (semantic): contracts (data contracts and product interface agreements), governance (policies, controls, stewardship), storage (storage and scaling architectures). Namespaces (procedural): pipelines (pipeline and data-movement patterns).

Entity types: data-contract (enforceable schema + semantics + SLA agreements between producers and consumers), data-product, data-governance-policy, data-quality-rule, storage-architecture, pipeline-pattern, data-platform.

Traits applied: cited.

Discovery patterns: recognizes data contract mentions, governance terminology, pipeline and storage architecture patterns.

When to bind

Bind data-engineering when researching data platform architecture, data contract adoption, data governance frameworks, or modern data engineering tooling and practices.

Enable

jq '(.ontologies[] | select(.id=="data-engineering") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-data-engineering topics.
Extends engineering-base, which itself extends mif-base and shared-traits; resolve-ontology.sh walks the full chain fail-closed — a missing engineering-base target aborts corpus resolution.
Scoped strictly to data-specific entity types; version 0.2.0 is a clean break with no back-compat aliases; technology is inherited from mif-generic, security types from software-security, and regulation from regulatory-legal.

Goals

Provides vocabulary for data engineering: data contracts, data products, governance policies, data quality rules, storage architectures, pipeline patterns, and data platforms.
Resolves shared engineering supertypes (component, architectural-decision, design-pattern, delivery-metric, engineering-practice, process-discipline) transitively via engineering-base without re-declaration.
Enables recognition of data contract definitions, governance terminology, and pipeline/storage architecture patterns in research sources.
Typed findings validate fail-closed against MIF schema on binding.

engineering-base

Version: 0.1.0 | Kind: ontology (shared layer — not directly bindable)

Source: schemas/ontologies/engineering-base/

Purpose

A MIF-compliant intermediate layer between the domain-neutral generic core (mif-base / mif-generic / shared-traits) and the engineering DOMAIN packs. It declares the supertypes that recur across every engineering domain so the domains inherit them instead of each re-declaring its own copy. It is a layer, not a bindable domain pack: topics do not bind engineering-base directly — they bind a descendant (software-engineering, data-engineering, software-security), whose extends chain reaches here.

Domain

Shared engineering vocabulary — architecture, components, patterns, decisions, delivery metrics, practices, and disciplines.

Entities, relationships, and traits

Entity types: component, architectural-decision, design-pattern, delivery-metric, engineering-practice, process-discipline, plus the cross-cutting universals control, artifact, policy, provenance (recurring across security, data, and software — domain packs specialize them via subtype_of, e.g. software-security security-control is subtype_of: [control] — a subtype is substitutable for its supertype at a relationship endpoint, enforced by the concordance validator and gate_m22).

Relationships: depends_on, implements, governs (control/policy → component/artifact), attests (provenance → artifact), derived_from (artifact lineage).

Traits applied: versioned, documented, dated, cited.

Discovery patterns: recognizes named design / architectural patterns (Factory, Singleton, Observer, Repository, Strategy, Decorator, CQRS, Event Sourcing, Saga).

Constraints

Not directly bindable; cataloged core=false — topics bind a descendant pack (software-engineering, data-engineering, or software-security), never this layer directly.
Resolved transitively only: resolve-ontology.sh walks the extends chain from a bound descendant; this layer is never itself the binding target.
Extends mif-base and shared-traits; the full chain is fail-closed — a missing or mistyped extends target in any descendant pack aborts corpus resolution.

Goals

Provides the shared engineering supertypes inherited by all engineering domain packs: component, architectural-decision, design-pattern, delivery-metric, engineering-practice, process-discipline, and the cross-cutting universals control, artifact, policy, and provenance.
Eliminates redundant supertype declarations across engineering domain packs; descendant packs declare only their domain-specific types and resolve supertypes via the extends chain.
Enables cross-pack subtype substitution: domain subtypes (e.g. security-control in software-security) are subtype_of these supertypes and substitutable at relationship endpoints, enforced by the concordance validator and gate_m22.
Recognizes named design and architectural patterns (Factory, Singleton, CQRS, Event Sourcing, Saga) in research sources via its discovery patterns.

Resolution

engineering-base is cataloged present-but-NOT-core (core=false): it is never always-on and never auto-applied to a non-engineering topic (biology, agriculture, legal never resolve these types). Resolution is transitive — binding a descendant pack resolves the supertypes this layer declares, because resolve-ontology.sh walks the extends chain. There is no Enable command and no “When to bind” step for this layer; enable and bind one of its descendant domain packs instead.

market-research

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/market-research/

Purpose

Provides an entity vocabulary for market and competitive research: market segments, competitors and brands, buyer personas, market sizing (TAM/SAM/SOM), competitive forces, service offerings and demand, value propositions, market-intelligence reports, data sources, survey instruments, and win-loss analyses. Sources include schema.org / GoodRelations, Umbrex market-mapping, HubSpot TAM/SAM/SOM, Porter Five Forces, and the Strategyzer Value Proposition Canvas.

Domain

Market analysis, competitive intelligence, and customer/segment research.

Entities, relationships, and traits

Entity types: segment, competitor, brand, buyer-persona, respondent-segment, sizing-estimate, competitive-force, service-offering, market-demand, value-proposition, market-intelligence-report, market-data-source, survey-instrument, win-loss-analysis.

Key relationships: analyzes-competitor, based-on-source, covers-segment, item-offered, maintains-brand, operates-in, provides-service, targets-audience, tracks-competitive-force.

Traits applied: auditable, bounded, categorized, contactable, located, measured, owned, reviewed, scheduled, scored, seasonal, tagged, versioned.

Discovery patterns: recognizes segment/vertical, competitor, buyer-persona/ICP, TAM/SAM/SOM sizing, Porter five-forces, survey/conjoint/NPS, win-loss, and data-source terminology.

When to bind

Bind market-research when researching market landscapes, competitive intelligence, customer segmentation, market sizing, or buyer / voice-of-customer analysis.

Enable

jq '(.ontologies[] | select(.id=="market-research") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-market-research topics.
Extends mif-base v1.0.0 (compatible with shared-traits v1.0.0); binding is fail-closed — resolve-ontology.sh aborts the corpus if the extends target is missing or mistyped.
Scoped to market and competitive research; entity types do not apply to scientific, legal, or engineering topics.

Goals

Provides vocabulary for market and competitive research: market segments, competitors, brands, buyer personas, market sizing (TAM/SAM/SOM), competitive forces, service offerings, value propositions, market-intelligence reports, survey instruments, and win-loss analyses.
Enables recognition of segment/vertical, competitor, TAM/SAM/SOM, Porter five-forces, NPS/conjoint survey, win-loss, and data-source terminology in research sources.
Grounded in schema.org/GoodRelations, Porter Five Forces, Strategyzer Value Proposition Canvas, and Umbrex market-mapping; every entity type traces to a named source class.
Typed findings validate fail-closed against MIF schema on binding.

observability

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/observability/

Purpose

Provides an entity vocabulary for observability-platform research: observability services and the telemetry signals they emit, service-ownership registries, capability comparisons against a baseline, market positioning, time-stamped roadmap signals, and platform migration patterns. Types form an explicit IS-A tree rooted in the inherited generic and engineering supertypes. It extends both engineering-base (for delivery-metric and design-pattern) and mif-generic (for technology and concept).

Domain

Observability platforms, telemetry, and platform migration analysis (for example AWS-native observability versus Datadog).

Entities, relationships, and traits

Namespaces (semantic): services (observability services and the signals and registries they provide), analysis (capability comparisons and market positioning). Namespaces (episodic): roadmap (GA dates, launches, deprecations, predictions). Namespaces (procedural): migrations (platform migration patterns and case studies).

Entity types: observability-service, service-ownership-registry, telemetry-signal, capability-comparison, market-position, roadmap-signal, migration-pattern. Two abstract intermediate supertypes (observability-resource, observability-assessment) organize the IS-A tree and are not stamped on findings.

Key relationships: emits, compares, positions, advances, catalogs.

Discovery patterns: recognizes observability service names (CloudWatch, X-Ray, ADOT, OpenTelemetry, Managed Prometheus, Managed Grafana, Application Signals), service-catalog and developer-portal terminology, capability parity and gap language, migration case studies, market-position language (Gartner, Magic Quadrant), and roadmap signals (GA, re:Invent, end-of-support).

When to bind

Bind observability when researching observability platforms, telemetry signals, service-ownership and catalog tooling, capability gaps between observability vendors, or platform migration patterns.

Enable

jq '(.ontologies[] | select(.id=="observability") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-observability topics.
Extends engineering-base and mif-generic; resolve-ontology.sh walks the full chain fail-closed — a missing or mistyped extends target aborts corpus resolution.
Scoped to observability platforms and migration analysis; the two abstract supertypes are not directly stampable on findings.

Goals

Provides vocabulary for observability-platform research: services, telemetry signals, service-ownership registries, capability comparisons, market positions, roadmap signals, and migration patterns.
Resolves shared engineering supertypes (delivery-metric, design-pattern) transitively via engineering-base and the generic technology and concept via mif-generic without re-declaration.
Enables recognition of observability service names, telemetry and pillar terminology, capability-parity language, market-position and roadmap signals, and migration case studies in research sources.
Typed findings validate fail-closed against MIF schema on binding.

psycholinguistics

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/psycholinguistics/

Purpose

Provides an entity vocabulary for psycholinguistics and computational stylometry research: psycholinguistic constructs, stylometric features, psychometric indices, elicitation protocols, research instruments, linguistic frameworks, and references to existing frameworks. Supports voice-elicitation, personality-language mapping, and authorship/readability research. Sources include the Big Five/OCEAN model, LIWC, Flesch readability, MATTR/MTLD lexical-diversity measures, Burrows’s Delta (stylo), and the Cognitive Interview.

Domain

Psycholinguistics, computational stylometry, and voice-elicitation research.

Entities, relationships, and traits

Namespaces (semantic): constructs (psycholinguistic and psychological constructs), features (stylometric and linguistic features), indices (psychometric indices and derived scores), instruments (research instruments, tools, and platforms), frameworks (linguistic and psycholinguistic frameworks), protocols (elicitation and interview protocols).

Entity types: psycholinguistic-construct, stylometric-feature, psychometric-index, elicitation-protocol, research-instrument, linguistic-framework, existing-framework-reference.

Key relationships: measures, operationalizes, grounds.

Traits applied: cited.

Discovery patterns: recognizes Big Five/OCEAN and HEXACO trait terminology, LIWC, readability and lexical-diversity measures (TTR, MATTR, MTLD, Flesch, Kincaid), stylometry and authorship-attribution terms (stylo, Burrows’s Delta), and elicitation-protocol language (cognitive interview, think-aloud, voice interview).

When to bind

Bind psycholinguistics when researching personality-language mapping, voice elicitation, stylometry and authorship attribution, readability and lexical diversity, or the research instruments and frameworks used in those studies.

Enable

jq '(.ontologies[] | select(.id=="psycholinguistics") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-psycholinguistics topics.
Extends engineering-base and mif-generic; resolve-ontology.sh walks the full chain fail-closed — a missing or mistyped extends target aborts corpus resolution.
Scoped to psycholinguistics and stylometry research; entity types do not apply to engineering operational, legal, or market-research topics.

Goals

Provides vocabulary for psycholinguistics and stylometry research: constructs, stylometric features, psychometric indices, elicitation protocols, research instruments, linguistic frameworks, and existing-framework references.
Resolves the generic concept and technology supertypes via mif-generic and delivery-metric and design-pattern via engineering-base without re-declaration.
Enables recognition of Big Five/OCEAN and LIWC terminology, readability and lexical-diversity measures, stylometry and authorship-attribution terms, and elicitation-protocol language in research sources.
Grounded in the Big Five/OCEAN model, LIWC, Flesch readability, MATTR/MTLD, Burrows’s Delta, and the Cognitive Interview; relationships are RO/IAO grounded.
Typed findings validate fail-closed against MIF schema on binding.

regenerative-agriculture

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/regenerative-agriculture/

Purpose

Provides an entity vocabulary for regenerative farm business operations: land, livestock, supply chain, carbon markets, and certification bodies. Sources include the Rodale Institute ROC Standards, Soil & Climate Initiative Verification Framework v3.0 (2025), USDA NRCS Soil Health Principles, Rainforest Alliance Regenerative Agriculture Standard (2025), and FAO Agroecology Knowledge Hub.

Domain

Regenerative farm business operations — farm records, supply chain, carbon credit activities, and certification tracking (not research observations).

Entities, relationships, and traits

Namespaces (semantic): land (land parcels, fields, soil profiles), livestock (animals, herds, breeding records). Additional namespaces cover supply chain, carbon markets, certifications, and farm financials.

Traits applied: lifecycle, owned, renewable, auditable, inventoried.

Discovery patterns: recognizes farm operation terminology, soil health references, certification body names (ROC, Rainforest Alliance), carbon market identifiers.

When to bind

Bind regenerative-agriculture when researching farm business operations, regenerative agriculture supply chains, carbon credit markets, or agricultural certification programs. For research-oriented findings about farming practices rather than farm records, use regenerative-agriculture-research instead.

Enable

jq '(.ontologies[] | select(.id=="regenerative-agriculture") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-agriculture topics.
Extends mif-base v1.0.0 and shared-traits v1.0.0; binding is fail-closed — resolve-ontology.sh and validate-concordance.sh abort the corpus if either extends target is missing or mistyped.
Scoped strictly to farm business records, supply chain, carbon credits, and certification tracking — not research observations; for research-oriented findings use regenerative-agriculture-research instead.

Goals

Provides vocabulary for regenerative farm business operations: land parcels and soil profiles, livestock, supply chain, carbon market activities, certifications, and farm financials.
Enables recognition of farm operation terminology, soil health references, certification body names (ROC, Rainforest Alliance), and carbon market identifiers in research sources.
Grounded in Rodale Institute ROC Standards, Soil & Climate Initiative Verification Framework v3.0 (2025), USDA NRCS Soil Health Principles, and FAO Agroecology Knowledge Hub.
Typed findings validate fail-closed against MIF schema on binding.

regenerative-agriculture-research

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/regenerative-agriculture-research/

Purpose

Provides a research-oriented entity vocabulary for regenerative agriculture findings. Covers research observations about farming practices, infrastructure, funding, and technology — not farm records. Types form an IS-A tree rooted in the inherited generic and engineering supertypes. It extends both engineering-base (for engineering-practice, component, and policy) and mif-generic (for technology and concept).

Domain

Research findings about regenerative agriculture practices: husbandry, agronomy, farm infrastructure, funding programs, and farm technology.

Entities, relationships, and traits

Namespaces (semantic): husbandry (animal husbandry and livestock care knowledge), agronomy (grazing, soil, crop, and pasture practices), infrastructure (fencing, irrigation, IoT, networks), funding (grants, cost-share, and funding programs).

Entity types (21): husbandry-practice, agronomic-practice, farm-infrastructure, grant-program, equipment-or-input, plus the husbandry leaves parturition-protocol, neonatal-care, periparturient-nutrition, livestock-health-condition, and breed-characteristic; the farm-infrastructure leaves fencing-system, irrigation-system, connectivity-link, sensor-or-edge-node, power-system, and network-segment; the equipment leaves iot-device and fencing-component; the grant-program leaf conservation-cost-share; and the cross-cutting adoption-trend and compliance-regulation.

Key relationships: relates_to, supports, has_part, connects.

Traits applied: cited.

Discovery patterns: recognizes husbandry and agronomy terminology, farm infrastructure keywords, grant and funding program identifiers, IoT and technology references in a farm context.

When to bind

Bind regenerative-agriculture-research when researching regenerative farming practices, husbandry techniques, soil science, agronomy research, or agricultural grant programs. For farm business records and supply chain tracking, use regenerative-agriculture instead.

Enable

jq '(.ontologies[] | select(.id=="regenerative-agriculture-research") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-agriculture topics.
Extends mif-base v1.0.0; binding is fail-closed — resolve-ontology.sh aborts the corpus if the extends target is missing or mistyped.
Scoped to research observations about farming practices — not farm business records or supply chain tracking; for farm records use regenerative-agriculture instead.

Goals

Provides research-oriented vocabulary for regenerative agriculture findings: husbandry practices, agronomy (grazing, soil, crop, pasture), farm infrastructure (fencing, irrigation, IoT), funding programs, and cross-cutting technology and security research types.
Enables recognition of husbandry and agronomy terminology, farm infrastructure keywords, grant and funding program identifiers, and IoT/technology references in a farm research context.
Cross-cutting technology and security research types are included so topics spanning farm technology and infrastructure resolve without a separate pack.
Typed findings validate fail-closed against MIF schema on binding.

regulatory-legal

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/regulatory-legal/

Purpose

Provides an entity vocabulary for regulatory and legal research: legislative acts and treaties, obligations and rights, jurisdictions and authorities, contracts, licenses, court decisions, regulatory sanctions, compliance reporting, and control mappings. Grounded in LKIF-Core, FIBO (FBC / FND), ELI v1.5, Akoma Ntoso, and NIST OSCAL.

Domain

Law, regulation, compliance, and governance contexts.

Entities, relationships, and traits

Entity types: legal-act, treaty, obligation, legal-right, legal-capacity, jurisdiction, authority, legal-person, legal-role, contract, license-permit, control-mapping, court-decision, regulatory-sanction, legal-procedure, assessment, compliance-report.

Key relationships: amends, applies_in, cites, confers, governed_by, has_jurisdiction_in, imposes, regulates, satisfies, transposes.

Traits applied: auditable, bounded, categorized, contactable, lifecycle, located, owned, regulated, renewable, reviewed, scheduled, scored, tagged.

Discovery patterns: recognizes regulation citations (Regulation (EU), U.S.C., GDPR, HIPAA), deontic language (shall / must / prohibited), jurisdictions, regulators, control crosswalks (OLIR), ELI / Akoma Ntoso URIs, case citations (ECLI), and contract / legal-role terminology. control-mapping.control_ref bridges to the software-security pack’s security-control type.

When to bind

Bind regulatory-legal when researching laws and regulations, compliance obligations, legal instruments and case law, or control-to-obligation mappings.

Enable

jq '(.ontologies[] | select(.id=="regulatory-legal") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-legal topics.
Extends mif-base v1.0.0 and shared-traits v1.0.0; binding is fail-closed — resolve-ontology.sh and validate-concordance.sh abort the corpus if either extends target is missing or mistyped.
Scoped to law, regulation, compliance, and governance; control-mapping.control_ref bridges cross-pack to software-security’s security-control type but the types are not interchangeable across packs.

Goals

Provides vocabulary for regulatory and legal research: legislative acts, treaties, obligations, rights, jurisdictions, authorities, contracts, licenses, court decisions, sanctions, compliance reports, and control mappings.
Enables recognition of regulation citations (Regulation (EU), U.S.C., GDPR, HIPAA), deontic language (shall / must / prohibited), ELI/Akoma Ntoso URIs, ECLI case citations, and OLIR control crosswalks in research sources.
Grounded in LKIF-Core, FIBO (FBC/FND), ELI v1.5, Akoma Ntoso, and NIST OSCAL; every entity type traces to a named source vocabulary class.
Typed findings validate fail-closed against MIF schema on binding; control-mapping.control_ref provides a cross-pack bridge to software-security’s security-control type.

scientific

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/scientific/

Purpose

Provides an entity vocabulary for scientific research: studies and investigations, methods and protocol applications, samples and measurements, hypotheses, instruments, publications and funding, and datasets with their catalogs, distributions, services, and provenance. Grounded in OBO Foundry / OBI, IAO, COB, W3C DCAT 3, W3C PROV-O, and schema.org (OBO IRIs are OLS4/Ontobee-confirmed).

Domain

Scientific studies, research data management, and data provenance.

Entities, relationships, and traits

Entity types: study, research-investigation, cohort, method, protocol-application, sample-organism, measurement, hypothesis, research-instrument, research-publication, research-funding, dataset, data-distribution, data-service, dataset-series, data-catalog, data-provenance.

Key relationships: applies, catalogs, enrolls, funded_by, has_sample, measured_on, produces, reports_in, tests, uses_instrument, uses_method.

Traits applied: auditable, bounded, budgeted, categorized, inventoried, lifecycle, located, maintainable, measured, owned, quality_controlled, renewable, reviewed, scheduled, tagged.

Discovery patterns: recognizes study / trial / cohort, assay / protocol / method, sample / organism / tissue, measurement, hypothesis, instrument, DOI / preprint, grant, DCAT dataset, and PROV-O provenance terminology.

When to bind

Bind scientific when researching scientific studies, experimental methods, research data and catalogs, or data provenance and lineage.

Enable

jq '(.ontologies[] | select(.id=="scientific") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-scientific topics.
Extends mif-base v1.0.0 and shared-traits v1.0.0; binding is fail-closed — resolve-ontology.sh and validate-concordance.sh abort the corpus if either extends target is missing or mistyped.
Scoped to scientific studies, research data management, and data provenance; entity types do not apply to engineering operational, legal, or market-research topics.
OBO IRIs are OLS4/Ontobee-confirmed gate-corrected values; a finding whose ontology.id and resolved type do not align is a hard fail, not a fallback.

Goals

Provides vocabulary for scientific research: studies, investigations, cohorts, methods, protocol applications, samples, measurements, hypotheses, instruments, publications, funding, datasets, data distributions, data services, dataset series, catalogs, and data provenance.
Enables recognition of study/trial/cohort, assay/protocol/method, sample/organism/tissue, measurement, hypothesis, instrument, DOI/preprint, grant, DCAT dataset, and PROV-O provenance terminology in research sources.
Grounded in OBO Foundry/OBI, IAO, COB, W3C DCAT 3, W3C PROV-O, and schema.org; every entity type traces to a named source vocabulary.
Typed findings validate fail-closed against MIF schema on binding.

software-engineering

Version: 0.5.0 | Kind: ontology

Source: packs/ontologies/software-engineering/

Purpose

Provides an entity vocabulary for the SDLC-operational slice of software engineering: production incidents and operational procedures. It extends engineering-base, from which it inherits the shared engineering supertypes (component, architectural-decision, design-pattern, delivery-metric, engineering-practice, process-discipline); the generic technology comes from mif-generic. Security types (security-threat, security-framework, security-incident) live in the software-security pack, and regulation is modeled in regulatory-legal (subsumed by its legal-act / obligation). The former adoption-trend is gone — the trend-analysis pack’s trend is canonical.

Domain

Software development teams, software architecture research, and engineering process analysis.

Entities, relationships, and traits

Namespaces (procedural): deployments (deployment procedures, release processes).

Entity types: incident-report, runbook, deployment-procedure, migration-guide (the shared supertypes are inherited from engineering-base, not re-declared here).

Traits applied: versioned, dated, timeline, stakeholders.

Discovery patterns: recognizes incident / outage / postmortem / RCA, runbook / playbook / SOP, deployment / release, and migration / upgrade terminology in research sources.

When to bind

Bind software-engineering when researching production incidents and postmortems, operational runbooks, deployment and release procedures, or system migration plans. For the shared engineering supertypes (components, architecture, decisions, patterns), bind this or any sibling engineering pack — they are inherited from engineering-base.

Enable

jq '(.ontologies[] | select(.id=="software-engineering") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-software-engineering topics.
Extends engineering-base, which extends mif-base and shared-traits; resolve-ontology.sh walks the full chain fail-closed — a missing target in the chain aborts corpus resolution.
Scoped strictly to SDLC-operational types; version 0.5.0 is a clean break with no back-compat aliases; security types belong to software-security, regulation to regulatory-legal, and trend to trend-analysis.

Goals

Provides vocabulary for software engineering operations: incident reports, runbooks, deployment procedures, and migration guides.
Resolves shared engineering supertypes (component, architectural-decision, design-pattern, delivery-metric, engineering-practice, process-discipline) transitively via engineering-base without re-declaration.
Enables recognition of incident/outage/postmortem/RCA, runbook/playbook/SOP, deployment/release, and migration/upgrade terminology in research sources.
Typed findings validate fail-closed against MIF schema on binding.

software-security

Version: 0.2.0 | Kind: ontology

Source: packs/ontologies/software-security/

Purpose

Provides an entity vocabulary for the software-facing slice of security research: vulnerabilities and weaknesses, controls, threat actors, campaigns and tactics, indicators of compromise, malware, tools and infrastructure, threat-intelligence reports, supply-chain risk, policies, assessments, and POA&Ms. It extends engineering-base. Grounded in MITRE ATT&CK / CAPEC, CVE / CWE / NVD, NIST SP 800-53 / OSCAL / 800-161r1, STIX 2.1, OWASP, and VERIS. The security-threat, security-framework, and security-incident types live here, as SDLC-facing supertypes that the finer STIX / ATT&CK / CWE types (attack-tactic, weakness, vulnerability) refine.

Domain

Cybersecurity threat intelligence, vulnerability management, and security compliance.

Entities, relationships, and traits

Entity types: attack-tactic, attack-mitigation, malware, vulnerability, weakness, security-control, threat-actor, attack-campaign, indicator-of-compromise, security-infrastructure, security-tool, threat-intelligence-report, supply-chain-risk, security-policy, security-assessment, poam, security-threat, security-framework, security-incident.

Key relationships: attributed_to, categorizes, defines, documents, exploits, hosts, indicates, mitigates, mitigates_threat, realizes, tracks, uses.

Traits applied: auditable, categorized, certified, inventoried, located, measured, owned, quality_controlled, regulated, reviewed, scheduled, scored, tagged, versioned.

Discovery patterns: recognizes ATT&CK technique / CAPEC IDs, CVE / CWE ids, NIST control ids, framework names, breach / ransomware terms, threat-actor and campaign names, IOC / YARA / STIX / TLP markers, supply-chain / SBOM, pen-test / red-team, and malware family names.

When to bind

Bind software-security when researching threat intelligence, vulnerability and weakness analysis, security controls and frameworks, or security compliance and assessment.

Enable

jq '(.ontologies[] | select(.id=="software-security") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-security topics.
Extends engineering-base (which extends mif-base and shared-traits); resolve-ontology.sh walks the full chain fail-closed — a missing target in the chain aborts corpus resolution.
security-threat, security-framework, and security-incident are defined here as SDLC-facing supertypes; finer ATT&CK/CWE types refine them. A finding whose resolved type belongs to a different ontology than its pin names is a hard fail.

Goals

Provides vocabulary for cybersecurity research: attack tactics, mitigations, malware, vulnerabilities, weaknesses, security controls, threat actors, attack campaigns, indicators of compromise, security infrastructure, tools, threat-intelligence reports, supply-chain risk, policies, assessments, and POA&Ms.
Resolves shared engineering supertypes transitively via engineering-base; security-control is subtype_of: [control] and substitutable at relationship endpoints, enforced by the concordance validator and gate_m22.
Enables recognition of ATT&CK technique/CAPEC IDs, CVE/CWE IDs, NIST control IDs, breach/ransomware terms, IOC/YARA/STIX/TLP markers, and SBOM/supply-chain terminology in research sources.
Typed findings validate fail-closed; control-mapping.control_ref in regulatory-legal bridges to this pack’s security-control type.

trend-analysis

Version: 0.1.0 | Kind: ontology

Source: packs/ontologies/trend-analysis/

Purpose

Provides an entity vocabulary for strategic foresight and trend analysis: weak signals, drivers, trends and megatrends, emerging issues, wild cards, critical uncertainties, adoption curves, forecasts, scenarios, horizons, implications, visions, and roadmaps. Grounded in IFTF foresight, EU JRC / K4P, Sitra, Shell / GBN scenario planning, Rogers Diffusion of Innovations, Gartner Hype Cycle, Three Horizons, the Futures Wheel, and OECD-OPSI. (Six types the inventory based on analytical are remapped to the semantic root, since the mif-base cognitive triad has no _analytical root.) This pack’s trend is canonical, replacing the former adoption-trend (which has been removed).

Domain

Strategic foresight, futures studies, and technology / market trend analysis.

Entities, relationships, and traits

Entity types: signal, driver, trend, megatrend, emerging-issue, wild-card, critical-uncertainty, adoption-curve, forecast, scenario, horizon, implication, vision, roadmap.

Key relationships: constrains, generates, grounds, indicates, informs, intensifies, matures_into, operationalizes, placed_on, produces, specializes.

Traits applied: auditable, bounded, categorized, measured, owned, reviewed, scheduled, scored, tagged, versioned.

Discovery patterns: recognizes weak-signal, driver-of-change / STEEP, trend, hype-cycle / S-curve, forecast, megatrend, scenario, and wild-card / black-swan terminology.

When to bind

Bind trend-analysis when researching strategic foresight, emerging signals and megatrends, technology adoption and hype cycles, or scenario and roadmap planning.

Enable

jq '(.ontologies[] | select(.id=="trend-analysis") | .enabled) |= true' \
  harness.config.json > harness.config.tmp && mv harness.config.tmp harness.config.json

Constraints

Opt-in only; cataloged core=false — never auto-applied to non-foresight topics.
Extends mif-base v1.0.0 (compatible with shared-traits v1.0.0); binding is fail-closed — resolve-ontology.sh aborts the corpus if the extends target is missing or mistyped.
trend is the canonical generic trend type here; the former adoption-trend from software-engineering is removed and replaced by this pack’s trend (no back-compat alias).
Six entity types (adoption-curve, forecast, horizon, implication, scenario, vision) are remapped to the semantic base under the _semantic/foresight namespace tree; the mif-base cognitive triad has no _analytical root.

Goals

Provides vocabulary for strategic foresight and trend analysis: signals, drivers, trends, megatrends, emerging issues, wild cards, critical uncertainties, adoption curves, forecasts, scenarios, horizons, implications, visions, and roadmaps.
Enables recognition of weak-signal, STEEP driver-of-change, trend, hype-cycle/S-curve, forecast, megatrend, scenario, and wild-card/black-swan terminology in research sources.
Grounded in IFTF, EU JRC/K4P, Sitra, Shell/GBN scenario planning, Rogers Diffusion of Innovations, Gartner Hype Cycle, Three Horizons, Futures Wheel, and OECD-OPSI.
Typed findings validate fail-closed against MIF schema on binding.