@aveltens/pi-kit-linked-data

Pi Coding Agent skills and extensions to work with RDF, SPARQL and other Linked Data technologies

Package details

extensionskill

Install @aveltens/pi-kit-linked-data from npm and Pi will load the resources declared by the package manifest.

$ pi install npm:@aveltens/pi-kit-linked-data

Pi manifest JSON

{
  "extensions": [
    "extensions/"
  ],
  "skills": [
    "skills/"
  ]
}

Pi packages can execute code and influence agent behavior. Review the source before installing third-party packages.

Pi Coding Agent skills and extensions to work with RDF, SPARQL, and other Linked Data technologies.

Adds tools and matching skills to the Pi Coding Agent:

Tool	What it does
`sparql_query_files`	Run SPARQL 1.1 queries against local RDF/Turtle files
`sparql_query_endpoint`	Query any remote SPARQL endpoint by URL
`sparql_query_wikidata`	Query the Wikidata Knowledge Graph for real-world facts

Two discovery tools are also included: discover_rdf_files and discover_sparql_queries.

pi install npm:@aveltens/pi-kit-linked-data

To try the package without installing it permanently:

pi -e npm:@aveltens/pi-kit-linked-data

To install per-project (shared with your team via .pi/settings.json):

pi install -l npm:@aveltens/pi-kit-linked-data

Put some Turtle files in your workspace, then ask Pi in plain English:

"What classes are defined in my ontology?" "List all SKOS concepts with their labels." "Find all resources of type schema:Person."

Pi will run discover_rdf_files to locate the files and sparql_query_files to query them.

"Query DBpedia for population of Germany." "List the classes available at https://example.org/sparql."

Pi uses sparql_query_endpoint and handles PREFIX declarations automatically.

"Who were the presidents of France in the 20th century?" "What is the capital of Japan?" "List all Nobel Prize winners in Physics since 2010."

Pi uses sparql_query_wikidata, which pre-injects all standard Wikidata prefixes (wd:, wdt:, wikibase:, etc.) — no declarations needed.

Tool	Description
`sparql_query_files`	Executes SPARQL 1.1 (SELECT/ASK/CONSTRUCT/DESCRIBE) against local `.ttl`, `.rdf`, `.n3`, `.jsonld`, `.trig`, `.nq`, `.nt` files using Comunica
`sparql_query_endpoint`	Executes SPARQL 1.1 against any remote endpoint over HTTP
`sparql_query_wikidata`	Executes SPARQL 1.1 against the Wikidata Query Service with standard prefixes pre-injected
`discover_rdf_files`	Finds all RDF data files in the workspace
`discover_sparql_queries`	Finds all `.rq` / `.sparql` query files in the workspace

Command	Description
`/rdf-memory-explorer`	Starts a local HTTP server and prints the URL in the chat — open it in any browser to browse stores and run SPARQL queries visually. Works in containers and remote environments.

Skills guide Pi on when and how to use the tools:

sparql-query-files — semantic graph traversal, class/property discovery, reusing existing .rq files
sparql-query-endpoint — endpoint exploration, output format selection, PREFIX declarations
sparql-query-wikidata — entity/property model, label service patterns, multi-hop traversal

The package ships two command-line tools you can run directly from your project.

Add convenience scripts to your package.json:

{
  "query": "cd .pi/npm && npx pi-kit-query",
  "validate": "cd .pi/npm && npx pi-kit-validate"
}

Then use them like:

npm run query
npm run validate