Docs: Add agent-docs for LLM/AI context

[data-goblin]

Description of new changes

• Summary: Adds docs/agent-docs/ directory containing context files for LLM/AI agents using the fab CLI. These markdown files provide structured guidance, command references, and workflow examples optimized for AI assistants. They are structured according to the Claude Code skill format. While the skill format is proprietary, the use of it is not.

• Context: AI coding assistants (Claude, Copilot, etc.) benefit from structured context files when working with CLI tools. These docs should help agents understand fab CLI syntax, common workflows, etc. They're focused on my use-cases, tho.

• Dependencies: None. These are standalone markdown documentation files. I've provided them also in a Claude Code Plugin which you can find here, which also contains a small MCP server that wraps the CLI, and some helper scripts that extend its functionality.

Files added

• SKILL.md - Main context file with command reference, core concepts, and common workflows
• references/ - Detailed guides for specific topics:
  • quickstart.md - Copy-paste examples for common operations
  • reference.md - Full command reference
  • semantic-models.md - TMDL, DAX, refresh operations
  • notebooks.md - Job execution and scheduling
  • workspaces.md - Workspace management
  • reports.md - Report operations
  • fab-api.md - Direct API access patterns
  • querying-data.md - DAX and SQL endpoint queries
  • create-workspaces.md - Workspace creation patterns

[aliabufoul]

Can you please elaborate on how users will benefit from these references?
Trying to understand the E2E experience of users utilizing this reference for better agent experience. And how it relates to the plugin you shared in case we wanted to distribute this skill as a plugin for Claude?

[aliabufoul]

Thanks for the contribution!
I noticed that many of the added references include examples of CLI commands, which are already covered in the following folder:
https://github.com/microsoft/fabric-cli/tree/main/docs/examples
This folder provides a rich set of scenarios. Have you considered referencing these existing examples instead of duplicating them?
Doing so would:

• Reduce the maintenance overhead of keeping two versions in sync
• Ensure agent have access to a more comprehensive and up-to-date set of examples