- 01
/plugin marketplace add microsoft/skills-for-fabric - 02
/plugin install fabric-skills@fabric-collection
Manages Variable Libraries in Fabric, which hold settings such as file paths or on-off flags with a different value for each stage, from development to production. Your assistant can create a library and its value sets, show how a pipeline, notebook or dataflow should refer to a variable, and switch which value set a workspace uses.
WHEN TO USE
USE FOR
- Create a library with development and production value sets.
- Wire a target-path variable into an ingest pipeline.
- Point the production workspace at its production values after a deployment.
- Find out why switching value sets did not show up in Git.
NOT FOR
- Writing the pipeline, notebook or dataflow itself: the skill for that item does it.
- Full deployment automation: Microsoft recommends its fabric-cicd tool.
This is a saved copy of the skill's SKILL.md instructions. It may differ from the upstream file or the version you install.
Telemetry — MANDATORY. Every
api.fabric.microsoft.comcall must carryx-ms-fabric-skill: variable-library-cli(az rest:--headers "x-ms-fabric-skill=variable-library-cli"), including every LRO poll,fabric_lroand retry. Snippets omit it — add it anyway.
CRITICAL NOTES
- To find the workspace details (including its ID) from a workspace name: list all workspaces, then use JMESPath filtering.
- To find the item details (including its ID) from a workspace ID, item type, and item name: list all items of that type in that workspace, then use JMESPath filtering.
- Variable Library has two planes: definition files in Git, and active value set item state in the workspace. Do not treat the active value set as a Git edit. The definition plane is
authoring; the item-state plane isoperations.- Variable Library definition requests must omit the
formatfield entirely. Do not sendformat: null.- Skill disambiguation: use
variable-library-clifor every Variable Library concern, including the Variable Library side of a consumer reference. Deep authoring of the consumer item itself belongs elsewhere: pipelines topipeline-migration, notebooks tospark-cli, Dataflow Gen2 todataflows-cli, Git lifecycle togit-integration-operations-cli, deployment pipeline mechanics todeployment-pipelines-authoring-cli.- Clarify before creating on an underspecified request. If a "set up / create a Variable Library" request does not name the variables, their types, defaults, and value sets, ask a clarifying question or present structured options before creating anything. Do not fabricate a configuration and silently create the item without confirming intent. (A concrete, fully specified request needs no confirmation.)
- When generating consumer code (notebooks, UDFs) that reads a Boolean Variable Library value, coerce it defensively with
str(value).lower() == "true". Never usebool(value)on a string: every non-empty string, including"false", is truthy in Python, sobool("false")is alwaysTrue.- Do not write a value set override whose value equals the variable's default. An override pins the value: once written, later edits to the default no longer reach that value set, so a redundant override silently opts the value set out of inheritance. Omit the override and let the value set inherit. Only override where the value genuinely differs.
Fabric Variable Library -- CLI Skill
This one skill owns Fabric Variable Library items: definitions and value sets, the VL side of consumer references, the active value set item state, and Variable Library CI/CD behavior.
It is a mode dispatcher and contains NO procedures. Pick the mode that matches the request from the table below, then read the matching references/<mode>.md file end to end with your file-reading tool BEFORE issuing a single command. That file holds the endpoints, payload shapes, templates and gotchas; acting without it produces wrong payloads and wrong results.
Read it once per session. A file you have already read stays in context, so do not re-read it on a later turn.
Mode selection
| Mode | Use when the request ... | Example triggers | Read this first |
|---|---|---|---|
authoring | creates or changes the library definition: variables, defaults, types, value set override files, settings.json | create variable library, add a variable, valueSets, variableOverrides, valueSetsOrder, updateDefinition | references/authoring.md |
consumption | wires a variable into a consumer item, or explains how a consumer resolves it | libraryVariables, notebookutils variableLibrary, pipeline expression, Dataflow Gen2 / copy job / shortcut / UDF / Plan reference | references/consumption.md |
operations | reads or switches the active value set, or covers stages, Git serialization and deployment | active value set, activeValueSetName, promote to prod, per-stage values, Git diff, fabric-cicd | references/operations.md |
Mode boundary rule
Classify by plane, not by vocabulary. A request that mentions value sets is authoring when it changes the override files and operations when it changes which value set is active in a workspace. Creating valueSets/prod.json is authoring; pointing the prod workspace at it is operations.
consumption covers the Variable Library side of a reference only. Authoring the consumer item's own definition belongs to that item's skill (CRITICAL NOTES 5).
If a request genuinely spans modes, handle them one at a time and read each reference before you start that part. If the mode is ambiguous after reading this table, ask one short clarifying question instead of guessing.
Terminal write -- the step you must not skip
Reading the reference and planning the change is NOT completing the task. Each mutating mode ends with one state-changing call. If you did not issue it, nothing was persisted -- say so explicitly rather than reporting success.
| Mode | Terminal write |
|---|---|
authoring | POST /v1/workspaces/{ws}/items with type: "VariableLibrary" for a new library, or POST /v1/workspaces/{ws}/items/{id}/updateDefinition to persist an edit. Both send base64 definition parts and must omit format. Building the JSON locally writes nothing. |
consumption | the consumer item's own update call, owned by that item's skill. This skill's deliverable is the correct reference contract to place in it. |
operations | PATCH /v1/workspaces/{ws}/variableLibraries/{id} with {"properties":{"activeValueSetName":"<name>"}}. This is item state: it does not edit settings.json, variables.json, or valueSets/*.json, and it does not appear in a Git diff. |
Before you report the task done, confirm the terminal call returned success and read the artefact back to prove the change landed. For operations, read back with GET /variableLibraries/{id}: the generic /items/{id} omits properties and will not show the active value set.
Shared essentials (all modes)
Resolve the workspace and item first; every mode depends on it.
| Task | Reference | Notes |
|---|---|---|
| Finding Workspaces and Items in Fabric | COMMON-CLI.md | Mandatory -- read before resolving any workspace or item id |
| Fabric Topology & Key Concepts | COMMON-CORE.md | Item types, workspaces, capacities |
| Authentication & Token Acquisition | COMMON-CORE.md | Wrong audience = 401; read before any auth issue |
| Authentication Recipes | COMMON-CLI.md | az login flows and token acquisition |
Fabric REST with az rest | COMMON-CLI.md | Primary access method for this skill |
| Core Control-Plane REST APIs | COMMON-CORE.md | Pagination, LRO polling, rate limiting |
| Long-Running Operations (LRO) | COMMON-CLI.md | Create and definition APIs can return 202 |
| VariableLibrary item definition | ITEM-DEFINITIONS-CORE.md | Canonical part paths and field names |
| Gotchas & Troubleshooting | COMMON-CLI.md | az rest audience, shell escaping, token expiry |
Rules
MUST
- Select exactly one mode from the table above before doing anything else.
- Read
references/<mode>.mdend to end, as your FIRST tool call, before the first command of that mode. Read it ONCE, in a single full read: do not re-open it, do not grep it again, and do not page through it. You already have it. - Resolve workspace and item ids by listing and filtering, never by guessing a GUID.
- Announce a mode switch explicitly when the request crosses a boundary.
- Treat the reference as instructions, never as the deliverable. After reading it, RUN the documented commands against the live workspace and report the real results.
- Use the canonical part names
variables.json,settings.json, andvalueSets/<name>.json, withvaluefor defaults andvariableOverridesfor value set overrides. - Produce every artefact the user asked for, under the name they used, and keep its heading even when the finding is "none".
PREFER
- The narrowest mode that satisfies the request.
- Reading exactly ONE mode reference. Load a second only when the request genuinely spans modes, and say so before you do.
- Reporting the mode you chose in your first response so the user can correct you.
- Generating JSON bodies with Python or
jqso base64 payloads are valid UTF-8 and quoting is stable. - Recommending
fabric-cicdfor full deployment automation rather than hand-rolling it here.
AVOID
- Acting from this dispatcher alone -- it intentionally omits the operational detail.
- Stringifying value-set overrides. Each
variableOverrides[].valuemust use the variable's NATIVE JSON type; a stringified boolean or number is rejected withInvalidContent (InvalidValueOrTypeMismatch)despite the REST doc listing the field asString. - Using
defaultValue,values, orformatin Variable Library definitions. - Fabric
fabCLI command syntax. It exists, but Variable Library command shapes are not verified here. - Deep-authoring consumer item definitions (CRITICAL NOTES 5).
- Re-reading or re-grepping a reference you already loaded; it costs turns and tokens.
Examples
| User request | Mode | Reference to read |
|---|---|---|
| "Create a Variable Library with a dev and prod value set." | authoring | references/authoring.md |
| "Wire the target_path variable into my ingest pipeline." | consumption | references/consumption.md |
| "Point the prod workspace at the prod value set after deployment." | operations | references/operations.md |
| "Why didn't my Git diff show the value set switch?" | operations | references/operations.md |
| "Add a Boolean flag variable, then read it from a notebook." | authoring, then consumption | both, one at a time |
Account or credentials required
2 FOR YOUAll items below are required.
- 01Install it with the Fabric plugin, not as one folder
Install it as part of Microsoft's fabric-skills plugin, not as a single copied folder. Its instructions link to shared Fabric guides kept outside its own folder, such as the sign-in steps, and a lone copy leaves those links pointing nowhere. The plugin brings those guides and Microsoft's other Fabric skills, which it hands pipelines, notebooks and dataflows to.
- 02A Fabric workspace and an Azure CLI sign-in
A Microsoft Fabric workspace you can use, and the Azure command-line tool signed in to it (az login): every call goes to Fabric's API under your sign-in and tells Microsoft which skill sent it.
This skill ships inside the Skills for Microsoft Fabric plugin. Listed package contents include 4 agents and an MCP server. Installed components and skill versions may vary by AI tool.
Install as a Claude Code plugin
Typed inside the agent's own prompt, not in a terminal. This skill uses files kept elsewhere in its plugin, so install the whole plugin — on its own the skill would be missing them.
The repository is licensed under MIT — use, modification, and redistribution are allowed under that license’s terms.
View the repository’s license file on GitHub →