Skill Claude CodeCodex
Load data into MarkLogic in bulk via the Flux pipeline — from an HTTP/HTTPS URL, local or S3 files, a JDBC database, or RDF files — and reprocess existing documents through server-side transform modules. Use when importing CSV, TSV, JSON, JSONL, Parquet, Avro, ORC, MLCP archives, or RDF; when loading open-data portals…
Skill Claude CodeCodex
Design MarkLogic multi-model data architecture — documents, RDF triples, and vectors — plus document URI schemes and the envelope pattern for data integration. Use when modelling a new domain, deciding whether relationships need triples or just nesting, laying out collections and URIs, designing TDE views, harmonizing…
Skill Claude CodeCodex
Build a MarkLogic FastTrack search UI — designing the search options set that drives facets, timelines, and maps, and scaffolding the React app that consumes it. Use when configuring search options for a faceted UI, adding facet or date-bucket or geospatial constraints, deciding between path-index and json-property…
Skill Claude CodeCodex
Configure OAuth2/OIDC Bearer token authentication on a MarkLogic app server, including external security objects, JWT claim to role or user mapping, app server settings across cluster groups, and the MCP server's oauth auth mode. Use when enabling JWT/OIDC login, when tokens authenticate but get no roles (HTTP 403…
Skill Claude CodeCodex
Diagnose and fix MarkLogic performance problems — slow searches, slow Optic or SPARQL queries, timeouts, high memory use, and ingest or merge trouble. Covers the E-node/D-node split, the two-step filtered-search process, cache interpretation, when a range index is mandatory, Optic plan reading, SPARQL join cost, and…
Skill Claude CodeCodex
Scaffold a deployable, source-controlled MarkLogic project using ml-gradle. Use when the goal implies something repeatable, version-controlled, or deployed to another environment — building an app, API, service, or backend; creating a new project or repo; adding a REST endpoint, transform, or resource extension…
Skill Claude CodeCodex
Choose and write the right MarkLogic query for a goal — full-text search, exact-value filtering, structured cts queries, string-grammar (cts.parse) queries, QBE, facets and distinct-value counts, Optic row queries over TDE views, SPARQL over the triple store, and vector similarity search. Use when composing any query…
Skill Claude CodeCodex
Design and build Retrieval-Augmented Generation pipelines on MarkLogic 12 as vector store and/or knowledge graph. Covers the three composable paradigms — Lexical RAG (BM25, no embeddings), Vector RAG (cosine, ANN top-k, and ANN+BM25 hybrid via vec.vectorScore), and Graph RAG (Semaphore concept classification scoping…
Skill Claude CodeCodex
Write server-side MarkLogic code — SJS and XQuery modules, REST resource extensions, Content Transformation Framework transforms, Flux reader and transform modules, and TDE templates — and apply MarkLogic application coding practices to them. Use when generating or debugging a module to deploy to the Modules database…
Skill Claude CodeCodex
Router for working with MarkLogic through this MCP server — maps a user goal to the MarkLogic-native capability and the tools that implement it. Use at the start of any MarkLogic task when the right approach is not obvious: loading or importing data, searching, analytics and aggregation, graph and RDF work…
Skill Claude CodeCodex
Diagnose and fix Semaphore content-classification quality problems in MarkLogic — false positives, missed matches, uniform or low scores, hierarchy not propagating, nearlist noise, poor results on short text, ignored title zones, and over-firing associative links. Covers the three fix levels (concept labels, classify…
Skill Claude CodeCodex
Architect a Semaphore + MarkLogic content-classification integration — choosing between inline classification at ingest, post-ingest reprocess enrichment, a REST transform, or a full Data Hub Framework pipeline. Covers CLS and KMM configuration and connectivity checks, the canonical document model for storing…
Skill Claude CodeCodex
Author, load, validate, and publish SKOS taxonomies in Semaphore KMM for use with MarkLogic content classification. Use when creating a new taxonomy or concept scheme, writing or editing SKOS Turtle, loading plain SKOS vocabularies (UNESCO, EuroVoc, AGROVOC, IPTC), fixing "No preferred labels" in Semaphore Studio…
Instructions file
Instructions for tternquist/marklogic-mcp, covering marklogic mcp — agent working instructions, core principle: problem-first thinking, where guidance lives — skills first, agent skills (.claude/skills/ /skill.md) and marklogic://instructions resource (src/resources/index.ts).
MCP server Claude CodeCodexCursor +2
MCP server "marklogic-mcp" as configured in tternquist/marklogic-mcp. Runs locally from the marklogic-mcp npm package.