agentleFS
santoshkanthetysantoshkanthety/powerbi-agent/skills/powerbi-te-cli/SKILL.md

powerbi-te-cli

Skill3 starsChanged 7 months agoMITView on GitHub

The cross-platform Tabular Editor CLI (`te`) — a single self-contained binary for Windows, macOS and Linux that inspects, edits, validates, deploys, refreshes and tests Power BI and Analysis Services semantic models, with structured JSON/CSV/TMDL output and safe-by-default preview diffs. Distinct from Tabular Editor 2's TabularEditor.exe. Use when the user mentions: te CLI, Tabular Editor CLI, te auth login, te bpa run, te deploy, te refresh, te test run, te vertipaq, te diff, te script, cross-platform Tabular Editor, Tabular Editor on Mac or Linux, model CI on Linux, headless semantic model edit.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before deploying to a shared or production model:

    line 119
  2. Very specific:

    No credentials on the command line: Use te auth login or a profile.

    line 157
  3. Distinctive:

    If a command fails on permissions, report the missing right.

    line 160

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Tabular Editor CLI (te)

te is a single self-contained binary that runs on Windows, macOS and Linux, built on the same engine as Tabular Editor 3. It is the first way to do real semantic-model work — BPA, deploy, refresh, VertiPaq analysis, assertion tests — from a Linux CI runner without a Windows agent and without .NET Framework.

Three different products, do not confuse them:

ToolExecutablePlatformThis skill
Tabular Editor 2TabularEditor.exeWindows only, freeno — powerbi-te2-cli
Tabular Editor 3TabularEditor3.exeWindows GUI, licensedno — see powerbi-te-docs
Tabular Editor CLIteWindows / macOS / Linuxyes

Flags and scripts are not portable between them. A TE2 invocation (-A, -D, -S, -B) will not work here; te uses verb-based subcommands.

Status — check this first

te is in Limited Public Preview. During the preview no licence is required, only a Tabular Editor account — but the preview build stops working after 2026-10-31, and licensing will be required at general availability.

Consequences to state before anyone builds on it:

  • A pipeline that depends on te will break at the preview cutoff unless the binary is refreshed. Do not put it on the critical path of a release train without a fallback.
  • The fallback is powerbi-te2-cli (free, Windows, stable) or an XMLA/TOM path via pbi-agent model.
  • Pin the downloaded version and record it, so a CI failure is diagnosable as a version change rather than a model change.

Setup

  1. Register or sign in at tabulareditor.com, then download the binary for the platform.
  2. te auth login — interactive device-code sign-in to Power BI / Fabric.
  3. te --help and te <command> --help — do this before first use of any command.

The surface is moving during preview. Read the installed binary's help rather than trusting any written reference, including this one. Where help and this file disagree, help wins.

te --version                 # record this in the pipeline log
te auth login
te --help

Connection state: te profile manages named connection profiles, te session manages session state, te config sets CLI defaults. Set a profile once per environment rather than repeating connection strings — and keep profile secrets out of the repo.

Command families

Verified family names and representative commands; confirm exact arguments with --help.

FamilyCommandsUse for
Init / savete init, te save-ascreate, convert between .bim / TMDL / PBIP
Editte set, te add, te remove, te moveproperty and object changes without a GUI
Inspectte list, te find, te diff, te depsenumerate objects, locate usage, compare, trace dependencies
Analysis & qualityte validate, te bpa run, te util, te vertipaqvalidation, Best Practice Analyzer, VertiPaq analysis
Executionte query, te script, te macroDAX queries, C# scripts, saved macros
Deploy & refreshte deploy, te refreshpush a model, trigger a refresh
Testingte test runassertion tests against a model
Shellte interactive, te completionREPL, shell completions

Design properties that matter in practice

  • Structured output — JSON, CSV and TMDL. Parse JSON; never scrape the human-readable table. Pass the output format flag explicitly in scripts rather than relying on a default.
  • Non-interactive mode — a global flag suppresses prompts. Set it in every CI invocation. A command that blocks on a prompt in a pipeline hangs until the job times out.
  • Safe by default — edits preview a diff before applying. In CI you will be disabling that confirmation; that makes reviewing the diff locally first a requirement, not a nicety.
  • Clear errors and exit codes — errors go to stderr with predictable exit codes. Check the exit code. Do not infer success from empty stderr or from the absence of the word "error".

Workflows

BPA in CI, on Linux

te auth login                              # or a profile with CI credentials
te bpa run <model-or-connection> --help    # confirm flags for this build
# then: run BPA, emit JSON, fail the job on any error-severity finding

Rules for the gate: fail only on the severities the team agreed to fail on, emit the full finding list as a build artifact, and keep the rule file in the repo under version control so a new finding is traceable to a rule change. Rule authoring lives in powerbi-bpa-rules.

Review a change before it reaches the model

te diff <source> <target>     # what this deploy would actually change
te deps <object>              # what breaks if this object changes
te validate <model>

te diff before te deploy is the habit to build. It converts "deploy and hope" into a reviewable change set, and it catches the two failures that hurt most: an unintended property reset, and a rename that orphans a report visual.

Deploy

te diff <src> <target>        # 1. always
te deploy <src> <target> --help

Before deploying to a shared or production model:

  • Show the user the diff summary — objects added, removed, changed — and get an explicit yes.
  • Never deploy over a model whose diff includes removals you cannot explain. A removal in a deploy diff is usually a stale source, not an intended deletion.
  • Deploy to a dev workspace first where one exists. Partial deployment options (what gets overwritten vs preserved: roles, partitions, data sources) differ by command — read --help and state which you are using.

Refresh and test

te refresh <model> --help     # confirm scope: full, table, partition
te test run <suite>

A refresh draws capacity. On a shared or production capacity, state the expected CU cost first — powerbi-fabric-capacity. Refresh strategy itself (incremental, partitions, policy) is powerbi-refresh-semantic-model.

VertiPaq analysis

te vertipaq <model> --help

This is the measurement that makes powerbi-performance-scale and powerbi-dax-performance actionable: column cardinality, dictionary and hierarchy sizes, compression, and which columns are actually costing memory. Capture it before and after an optimisation and report both numbers.

Guardrails

  • Discovery before invocation. te <command> --help on first use in a session. The preview surface changes between builds.
  • Never run an edit, deploy or refresh against a production model without naming the model and the change and getting a yes. te is headless; there is no GUI prompt to catch a mistake.
  • No credentials on the command line. Use te auth login or a profile; command-line arguments are visible to other processes on the host.
  • Record the binary version in any pipeline that uses it, given the preview cutoff.
  • If a command fails on permissions, report the missing right. Do not retry through a different tool to get around it.

Related skills

NeedSkill
Windows-only, free, stable CLIpowerbi-te2-cli
Tabular Editor documentation lookup, scripting cookbookpowerbi-te-docs
C# scripts the CLI executespowerbi-c-sharp-scripting
BPA rule authoringpowerbi-bpa-rules
What to do with VertiPaq findingspowerbi-performance-scale, powerbi-dax-performance
Refresh designpowerbi-refresh-semantic-model
CU cost of a refresh or deploypowerbi-fabric-capacity
TMDL the CLI reads and writespowerbi-tmdl

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.