specification: API Commons AgentSkills specificationVersion: '0.1' provider: CData providerId: cdata generated: '2026-09-05' method: searched source: >- https://docs.cloud.cdata.com/.well-known/agent-skills/cdata/skill.md and https://github.com/CDataSoftware/connect-ai-skills (raw.githubusercontent.com/ CDataSoftware/connect-ai-skills/main/...). All three files below are CData's own, saved VERBATIM. Nothing in this directory was authored by API Evangelist. note: >- CData publishes Agent Skills as a product surface, not as an afterthought: a discovery skill served from the docs host's /.well-known/agent-skills/ path (named by the A2A agent card), plus a public repository of 20+ MIT-licensed skills installable with `npx skills add CDataSoftware/connect-ai-skills`. A second repository, CDataSoftware/cli-skills, covers the CData CLI. Because the provider publishes its own, no skills were generated on CData's behalf. skills: - id: cdata name: Cdata file: cdata-connect-ai-published-skill.md method: searched source: https://docs.cloud.cdata.com/.well-known/agent-skills/cdata/skill.md http_status: 200 content_type: text/markdown bytes: 10629 api: CData Connect AI description: >- The discovery skill the A2A agent card points at. Product summary, when-to-use triggers, connection status states, permission types, system roles, and the endpoint map (REST https://cloud.cdata.com/api/query, OData https://cloud.cdata.com/api/odata/{workspace_name}, SQL Server tds.cdata.com:14333, MCP https://mcp.cloud.cdata.com/mcp). frontmatter: name: Cdata metadata: {mintlify-proj: cdata, version: '1.0'} - id: connect-ai-base name: CData Connect AI — Base Skill file: cdata-connect-ai-base.md method: searched source: https://raw.githubusercontent.com/CDataSoftware/connect-ai-skills/main/skills/connect-ai-base/SKILL.md http_status: 200 bytes: 14430 license: MIT api: CData Connect AI MCP operations: - getInstructions - getCatalogs - getSchemas - getTables - getColumns - queryData description: >- The required base skill — every connector-specific skill in the repository composes on top of it. Encodes the mandatory discovery workflow (getInstructions before any schema/table/column call), tool-availability confirmation via tool_search, query construction patterns, SQL dialect guidance, and error recovery. - id: connect-ai-onboarding-claude-code name: Connect AI onboarding for Claude Code file: cdata-connect-ai-onboarding-claude-code.md method: searched source: https://raw.githubusercontent.com/CDataSoftware/connect-ai-skills/main/skills/onboarding/connect-ai-onboarding-claude-code/SKILL.md http_status: 200 bytes: 22899 license: MIT api: CData Connect AI MCP description: >- Guided first-time setup for Claude Code: create a Connect AI account, authenticate to the MCP server (PAT by default, browser OAuth alternative), connect a data source, run a first query. published_elsewhere: repository: https://github.com/CDataSoftware/connect-ai-skills install: npx skills add CDataSoftware/connect-ai-skills connector_skills_listed_in_repo_readme: 18 families: [CRM, ERP, HCM, Accounting, Collaboration, Files, Marketing] named: - connect-ai-salesforce - connect-ai-bullhorncrm - connect-ai-dynamics365businesscentral - connect-ai-workday - connect-ai-quickbooksonline - connect-ai-sageintacct - connect-ai-stripe - connect-ai-jira - connect-ai-confluence - connect-ai-airtable - connect-ai-docusign - connect-ai-googlecalendar - connect-ai-googlesheets - connect-ai-exchange - connect-ai-gmail - connect-ai-googledrive - connect-ai-instagram note: >- Only the base and onboarding skills are mirrored here. The connector-specific skills describe THIRD-PARTY data models (Salesforce, Workday, Stripe…) reached through Connect AI, so they belong to those vendors' domains as much as to CData's; the repository is linked rather than copied. cli_skills_repository: https://github.com/CDataSoftware/cli-skills cli_install: npx skills add CDataSoftware/cli-skills maintainers: - FN: Kin Lane email: info@apievangelist.com