BetaThis is a Beta (test) version. Content is being refined; the official release will prevail.
Global AI Agents · Hainan Hub Hotline 18913116626 / 13307508848
HNAIA Hainan Free Trade Port AI Agent Association (Preparatory) Hainan Free Trade Port AI Agent Association (Preparatory)Hainan Free Trade Port AI Agent Association (Preparatory)
Home/Skill Ecosystem/Technical Standards

Technical Standards

Skill Encapsulation Technical Standard (Draft) · Open for Comments

Skill Encapsulation Technical Standard (Draft)

Association Group Standard · Open for Comments

⬇ Download Full Standard (PDF)

Skill is the smallest tradable unit of wisdom in the association ecosystem. To ensure Skills are composable, verifiable, priceable and traceable, the Association formulates unified encapsulation standards covering structure, interfaces, dependencies, testing, ethics declarations and compute labels.

I. Basic Skill Structure

ComponentDescriptionFormat Requirement
Skill Description FileMetadataJSON-LD, following Schema.org standards
Interface DefinitionInput parameters, output results, error codesOpenAPI 3.0 specification
Business Logic CodeCore algorithms, model weightsPython 3.9+ / Node.js 18+
Dependency ListThird-party libraries, datasetsrequirements.txt / package.json
Test CasesUnit tests, integration tests, boundary testspytest / Jest
Ethics DeclarationAlgorithm fairness, data privacyAssociation unified template
Compute Demand LabelCPU / GPU consumption, energy gradeA–G unified energy efficiency rating

II. Compute-Power Synergy Adaptation Standard (Hainan Feature)

  • Must declare compute demand grade (A–G) as a basic parameter for scheduling and billing.
  • Support dynamic scheduling interface to receive grid load signals and adjust computing timing.
  • Prioritize use of clean energy labels and participate in green power direct-supply matching.
  • Skills running at Lingshui Undersea Data Center automatically receive the green power certification label.

III. Security Classification

GradeDefinitionApplicable ScenariosReview Requirements
L1 PublicGeneral tools, no sensitive data processingTranslation, calculators, format conversionAutomated detection
L2 FilingProcesses commercial dataSmart customer service, marketing copyFormal review + sampling test
L3 ReviewAffects users\' substantive rightsFinancial recommendations, medical consultationSubstantive review + expert evaluation
L4 LicensedInvolves public safetyAutonomous driving, diagnosis and treatment, low-altitude airspaceGovernment license + continuous supervision

IV. Skill Description File Example

// skill.jsonld — Skill metadata description file (excerpt)
{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "Aerospace Materials Multi-Physics Simulation",
  "identifier": "hnaia:skill:2026:000137",
  "provider": { "@type": "Organization", "name": "Hainan Commercial Launch Agent Lab" },
  "hnaia:securityLevel": "L3",
  "hnaia:energyGrade": "A",
  "hnaia:greenCertified": true,
  "hnaia:crossBorderReady": true,
  "hnaia:guardianMemberId": "HNAIA-P-2026-0042",
  "offers": { "@type": "Offer", "priceCurrency": "CNY", "price": "2800", "hnaia:pricingModel": "per-call" },
  "hnaia:publicWelfareRate": "0.05"
}
📝

Feedback

During the open comment period for the standard draft, members and industry institutions are welcome to submit revision suggestions to info@hnaia.org. Please indicate "Skill Standard Comment—Organization—Name" in the email subject.