gemini-enterprise-custom-connector-skill

How to build a custom connector for Gemini Enterprise to make a data source queryable that isn't in the native connector library (e.g., Snowflake, SAP, internal databases, proprietary REST APIs). Covers the two supported paths — Custom MCP Server (live tool calling) and Custom Connector → Discovery Engine datastore (indexed search/RAG) — when to use each, scaffolding, Python SDK calls, OAuth/ACL configuration, and registering with a Gemini Enterprise app. Use this skill whenever the user mentions Gemini Enterprise alongside any source that isn't in the natively supported list (Salesforce, ServiceNow, SharePoint, Jira, Confluence, Drive, Slack, GitHub, Notion, Box, Dropbox, etc.) — even without the words "custom connector." Also use it for "scaffold a connector," "make X searchable in Gemini Enterprise," "ingest X into a Gemini datastore," "expose X as a tool to a Gemini agent," or any question about Gemini Enterprise reading from Snowflake / SAP / a warehouse / an internal API.