robos:AgentSkill

Validation shape for robos:AgentSkill nodes in the learning package group.

Details

Class	robos:AgentSkill
Shape id	robos:AgentSkillShape
Package	eLearning Curriculums (robos.learning) (learning)
Classification	Agents & MCP
Schema.org type	schema:HowTo

Properties

Count is the minimum and maximum number of values (* means no maximum). The validator enforces both. The type column is documentation only; the validator does not check types.

Property	Count	Type	Rule or description
dcterms:title	1..*	xsd:string	Agent skill needs a title.
robos:skillName	1..*	-	Agent skill needs its declared name.
robos:sourcePath	1..*	-	Agent skill needs its instruction file.
robos:inRepository	1..*	-	Agent skill needs its repository reference.
robos:evidence	1..*	-	Agent skill needs source evidence.

Example node

This node comes from .robos/kgraphs/learning/package.jsonld in the RobOS repository. That graph holds sample data for a fictional company, so treat the values as illustrations.

{
  "@id": "urn:robos:source:import-skill",
  "@type": [
    "robos:AgentSkill",
    "schema:HowTo"
  ],
  "dcterms:title": "Evidence-backed graph import skill",
  "robos:package": "learning",
  "robos:sourcePath": "plugins/robos/skills/import-company-kgraph/SKILL.md",
  "robos:sourceKind": "implementation",
  "robos:inRepository": {
    "@id": "urn:robos:source:workspace"
  },
  "robos:evidence": [
    {
      "repository": "robos",
      "path": "plugins/robos/skills/import-company-kgraph/SKILL.md",
      "line": 1
    }
  ],
  "robos:skillName": "import-company-kgraph",
  "robos:uses": {
    "@id": "urn:robos:source:source-extractor"
  },
  "robos:namespace": "robos.learning"
}

Validate the example

Save the example as node.json in the repository root, then run:

node -e "const { SHACLValidator } = require('./packages/robos-graph/lib/shacl-validator'); console.log(new SHACLValidator().validate([require('./node.json')]).conforms)"

The command prints true when the node meets every rule for its types.