oracle ai data platform workbench spark connectors — Agent Skill

Oracle AI Data Platform Workbench — Spark Connectors

Canonical home: `oracle-samples/oracle-aidp-samples/ai/claude-code-plugins/oracle-ai-data-platform-workbench-spark-connectors`. This repository is now a personal development mirror. End users should install via Anthropic's community marketplace (see below), which sources from the canonical Oracle-org location.

A Claude Code plugin that ships 27 model-invokable skills for connecting Oracle AI Data Platform Workbench Apache Spark notebooks to every data source these notebooks commonly need. Each skill produces plain Python (Spark JDBC, Spark structured streaming, Spark oci:///s3a:///abfss://, or REST → Spark DataFrame) that runs in the notebook without any additional runtime.

v0.8.0 adds Google BigQuery guidance for AIDP 5.1: service-account authentication, ingestion reads, SQL pushdown, catalog.id, and external-catalog reads. GBQ is read-only.

All connectors wrap the official AIDP aidataplatform Spark format handler (or, where applicable, Spark JDBC / structured streaming / oci:///s3a:///abfss://) — same patterns shown in the upstream `oracle-samples/oracle-aidp-samples` connector notebooks.

Install

From Anthropic's community plugin marketplace:

/plugin marketplace add anthropics/claude-plugins-community
/plugin install oracle-ai-data-platform-workbench-spark-connectors

What's in here

29 skills total (27 connectors + 1 bootstrap + 1 routing).

Oracle / OCI sources

| Skill | Target | Transport | Recommended auth | |---|---|---|---| | aidp-connectors-overview | (router) | — | — | | aidp-connectors-bootstrap | one-time setup | — | — | | aidp-alh | Oracle Autonomous DB family (ALH / ADW / ATP) | Spark JDBC | Wallet (mTLS), IAM DB-Token, or API key | | aidp-oracle-db | Generic Oracle DB (Compute / Base DB / on-prem / 19c-26ai non-Autonomous) | aidataplatform (type=ORACLE_DB) | Plain user/password | | aidp-exacs | Exadata Cloud Service | Spark JDBC (TCP 1521 + NNE AES256) | Plain user/password | | aidp-peoplesoft | Oracle PeopleSoft | aidataplatform (type=ORACLE_PEOPLESOFT) | Plain user/password (read-only) | | aidp-siebel | Oracle Siebel CRM | aidataplatform (type=ORACLE_SIEBEL) | Plain user/password (read-only) | | aidp-fusion-rest | Fusion ERP/HCM/SCM | REST → DataFrame | HTTP Basic | | aidp-fusion-bicc | Fusion BICC bulk extracts | aidataplatform (type=FUSION_BICC) | HTTP Basic | | aidp-epm-cloud | EPM Cloud Planning | REST → DataFrame | HTTP Basic (tenancy.user@domain) | | aidp-essbase | Essbase 21c | REST + MDX → DataFrame | HTTP Basic | | aidp-streaming-kafka | OCI Streaming | Spark structured streaming | SASL/PLAIN with OCI auth token | | aidp-object-storage | OCI Object Storage native | Spark oci:// | Implicit IAM (workspace identity) | | aidp-iceberg | Apache Iceberg on OCI Object Storage | Iceberg Hadoop catalog on oci:// | Implicit IAM |

External RDBMS / Hadoop (aidataplatform format)

| Skill | Target | Transport | Recommended auth | |---|---|---|---| | aidp-postgresql | PostgreSQL | Spark JDBC (runtime-loaded driver) for SSL targets, aidataplatform type=POSTGRESQL for non-SSL | Plain user/password | | aidp-mysql | MySQL / OCI MySQL HeatWave | aidataplatform (type=MYSQL or MYSQL_HEATWAVE) | Plain user/password | | aidp-sqlserver | Microsoft SQL Server | aidataplatform (type=SQLSERVER) | Plain user/password | | aidp-azuresql | Azure SQL Database | aidataplatform (type=AZURE_SQLSERVER) | Plain user/password | | aidp-db2 | IBM Db2 | aidataplatform (type=DB2) | Plain user/password | | aidp-hive | Apache Hive (HiveServer2, non-Kerberos) | aidataplatform (type=HIVE) | Plain user/password |