Bundle 25 Databricks skills as coda-databricks-skills plugin

Summary

Bundles the field-engineering Databricks skill set as a CODA plugin, auto-loaded at startup via the marketplace. Pure addition — no production code changes. Migrating from datasciencemonkey PR #141.

25 skill domains synced from databricks-solutions/ai-dev-kit:
Agent Bricks, AI Functions, AI/BI Dashboards, Apps (Python), Asset Bundles, BDD Testing, Config, DBSQL, Docs, Execution Compute, Genie, Iceberg, Jobs, Lakebase (Autoscale + Provisioned), Metric Views, MLflow Evaluation, Model Serving, Python SDK, Spark SDP, Spark Structured Streaming, Synthetic Data Gen, Unity Catalog, Unstructured PDF, Vector Search, ZeroBus Ingest, Python Data Source.

Dependency

Depends on the plugin loader landing first (the sibling coda-essentials-plugin issue). Without that, this plugin doesn't get loaded.

Branch

feat/coda-databricks-skills — about to be pushed.

Diff scope

+51071 / 0, 156 files. 100% markdown + one plugin.json. No production code touched.

Open question

Vendor these skills here, or keep them refreshable from databricks-solutions/ai-dev-kit via a CI sync? Current design is "vendored at PR-merge time" — refreshes need a new PR. Worth deciding before merge.

Review the plugin loader dependency from the sibling coda-essentials-plugin issue, then inspect the planned feat/coda-databricks-skills branch and plugin.json. Verify that the 156 markdown files are bundled and auto-loaded through the marketplace. Before merge, resolve whether skills remain vendored from databricks-solutions/ai-dev-kit or use a CI refresh flow.