powerbi-desktop

name	powerbi-desktop
description	Use when working with Microsoft Power BI Desktop files or projects, including PBIX/PBIT inventory, DAX and Power Query review, semantic model documentation, Tabular Editor workflows, and local Desktop environment checks.

Power BI Desktop

Use this skill when the user asks Codex to inspect, document, generate, or refactor Power BI Desktop assets.

Boundaries

• Treat .pbix and .pbit files as package files. Do not modify them directly unless the user has exported a supported editable project format or provides explicit tooling.
• Prefer Power BI Project folders (.pbip), Tabular Model Definition Language (.tmdl), model.bim, Power Query .pq, DAX .dax, JSON metadata, and exported text over binary PBIX edits.
• Before editing, identify whether the workspace contains a PBIP folder, extracted metadata, or only binary PBIX/PBIT files.
• Do not publish to Power BI Service, sign in, refresh credentials, or overwrite reports unless the user explicitly asks and the required CLI/tooling is available.

First Checks

1. Inventory the workspace with scripts/Get-PowerBIInventory.ps1.
2. Check local Desktop/tooling availability with scripts/Test-PowerBIEnvironment.ps1.
3. If Power BI Desktop is open, detect the live model with scripts/Get-PowerBIDesktopLiveConnection.ps1.
4. Analyze PBIP structure with scripts/Get-PowerBIPBIPStructure.ps1.
5. For full AI review, run scripts/Invoke-PowerBIAutoReview.ps1.
6. Run an insight scan with scripts/Invoke-PowerBIInsightScan.ps1.
7. Generate a metric catalog with scripts/New-PowerBIMetricCatalog.ps1 when DAX files are available.
8. Generate a dependency graph with scripts/New-PowerBIDependencyGraph.ps1 when measures exist.
9. Generate an AI prompt pack with scripts/New-PowerBIAIPromptPack.ps1 for deeper Codex/LLM review.
10. Generate a refactoring plan with scripts/New-PowerBIRefactorPlan.ps1 when findings exist.
11. Generate a report blueprint with scripts/New-PowerBIReportBlueprint.ps1 when measures exist.
12. If text-based model files exist, summarize them with scripts/New-PowerBIModelSummary.ps1.
13. For advanced remediation, run scripts/Invoke-PowerBIInnovationReview.ps1.
14. For safe authoring, generate drafts with scripts/New-PowerBIMeasureDraft.ps1 or scripts/New-PowerBICalculatedColumnDraft.ps1.
15. For report authoring, create PBIP page/visual drafts with scripts/Add-PowerBIPBIPReportPage.ps1.
16. For native external-tool-style capabilities, run scripts/Invoke-PowerBINativeToolParityReview.ps1.
17. For real feature coverage, run scripts/Invoke-PowerBIRealFeatureReview.ps1.
18. For applying PBIP-safe drafts, use scripts/Apply-PowerBIPBIPMeasureDraft.ps1, scripts/Apply-PowerBIPBIPCalculatedColumnDraft.ps1, scripts/Apply-PowerBIPBIPPowerQueryDraft.ps1, or scripts/Apply-PowerBIPBIPTmdlDraft.ps1.
19. For optional external tool validation workflows, run scripts/Invoke-PowerBIExternalToolsReview.ps1.
20. For a single productized pass, run scripts/Invoke-PowerBIUnifiedReview.ps1.
21. For Power BI External Tools menu integration, generate a .pbitool.json with scripts/New-PowerBIExternalToolRegistration.ps1, or install/uninstall with scripts/Install-PowerBIExternalTool.ps1 and scripts/Uninstall-PowerBIExternalTool.ps1.
22. Before publishing plugin changes, run scripts\Test-PowerBIGoldenBaselines.ps1.
23. For maximum AI workflows, use scripts/Invoke-PowerBIAutonomousFixAgent.ps1, scripts/Compare-PowerBILiveRepoModel.ps1, scripts/New-PowerBIKpiTrustContract.ps1, scripts/Invoke-PowerBIAskModel.ps1, and scripts/New-PowerBIFabricReadinessPlan.ps1.
24. For the complete 39-artifact AI USP package, run scripts/Invoke-PowerBIMaxAIReview.ps1.
25. For enterprise release engineering, run scripts/New-PowerBIReleaseCandidatePack.ps1, scripts/New-PowerBIServiceScanner.ps1, scripts/New-PowerBIModelRiskHeatmap.ps1, scripts/Invoke-PowerBISemanticTestRunner.ps1, and scripts/New-PowerBITomWritePlan.ps1.
26. For local business process data quality checks across Power BI metadata and CSV/JSON exports, run scripts\New-PowerBIProcessDataMapping.ps1, scripts\Invoke-PowerBIBusinessProcessDataQuality.ps1, or scripts\New-PowerBIBusinessProcessDQPack.ps1.
27. For analytical release QA, run scripts\Test-PowerBIAnalysisMethodology.ps1, scripts\New-PowerBIMetricChangeDiagnosis.ps1, or scripts\New-PowerBIAnalyticalReleaseReport.ps1.
28. For advanced release USPs, run scripts\New-PowerBIEvidenceGraph.ps1, scripts\Test-PowerBISemanticContract.ps1, scripts\New-PowerBIExecutiveTrustBrief.ps1, or scripts\New-PowerBIReleaseCandidatePack.ps1 -IncludeAdvancedUspQa.
29. For separated portfolio, compliance, and operations QA, run scripts\New-PowerBIReleaseCandidatePack.ps1 -IncludePortfolioGovernanceQa -IncludeComplianceQa -IncludeOperationsQa.
30. For Fabric live read-only QA, run scripts\Get-PowerBIFabricAccessPlan.ps1, scripts\Import-PowerBIFabricWorkspaceSnapshot.ps1, or scripts\New-PowerBIReleaseCandidatePack.ps1 -IncludeFabricLiveQa -IncludeFabricPortfolioQa -IncludeFabricDeploymentQa -IncludeFabricOperationsQa -IncludeFabricGovernanceQa -IncludeFabricExecutiveQa.
31. When the user asks whether features are mocked, simulated, draft-only, or live-backed, run scripts\New-PowerBIFeatureMaturityMap.ps1.
32. For disruptive autonomous planning workflows, use the dedicated skills powerbi-autonomous-planning-loop, powerbi-goal-seeking-planning, powerbi-constraint-aware-planning, powerbi-autonomous-forecast-agents, powerbi-autonomous-exception-management, powerbi-revenue-rescue-mode, powerbi-forecast-trust-market, powerbi-causal-counterfactual-forecasting, powerbi-self-healing-forecast-governance, powerbi-planning-memory, powerbi-planning-readiness-score, and powerbi-forecast-war-room.
33. Look for these editable artifacts:

• *.pbip
• *.SemanticModel, *.Report
• definition.pbism, model.bim
• *.tmdl, *.dax, *.pq
• DiagramLayout, report.json, pages.json

Common Workflows

Inventory reports

Run:

.\plugins\powerbi-desktop\scripts\Get-PowerBIInventory.ps1 -Path .

Summarize report files, sizes, modification dates, and editable metadata candidates. If only PBIX/PBIT files exist, explain that binary-safe inspection is limited.

Document a model

• Prefer TMDL or model.bim as source.
• Extract tables, columns, measures, relationships, partitions, calculation groups, roles, and data sources.
• Keep DAX and M code in fenced code blocks when reporting findings.
• Flag measures with ambiguous names, hidden dependencies, hard-coded dates, repeated filter logic, or suspicious inactive relationship usage.

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIModelSummary.ps1 -Path . -OutputPath .\powerbi-model-summary.md

Use the generated markdown as a first-pass map. Verify important business logic against the source files before making recommendations.

Run an insight scan

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBIInsightScan.ps1 -Path . -OutputPath .\powerbi-insight-scan.md

Use this as the default "innovative" entry point. It combines file discovery, DAX risk heuristics, Power Query dependency checks, source-control readiness, and recommended next actions. Treat findings as triage signals, not proof of defects.

Run the full AI review package

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBIAutoReview.ps1 -Path . -OutputDirectory .\powerbi-auto-review

Use this as the fastest end-to-end path. It produces inventory, PBIP readiness, scan, metric catalog, dependency graph, refactor plan, report blueprint, model summary, executive narrative, AI prompt pack, and innovation review artifacts.

Run the unified product review

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBIUnifiedReview.ps1 -Path . -OutputDirectory .\powerbi-unified-review

Use this when the user wants the complete productized workflow. It runs the offline auto-review, attempts the live Desktop review when a local endpoint is available, generates the External Tools registration file, and writes a single index plus machine-readable summary.

Run the innovation review package

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBIInnovationReview.ps1 -Path . -OutputDirectory .\powerbi-innovation-review

Use this when the user wants actionable next-generation review artifacts. It produces a guided fix plan, semantic remediation guidance, measure lineage impact, generated DAX test plan, performance advisor, report UX critique, executive explainability pack, governance scorecard, Copilot readiness check, release checklist, business semantic layer, KPI trust score, decision-risk assistant, DAX fix simulation, visual-to-measure impact map, and trust release gate.

Run maximum AI engineering workflows

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBIAutonomousFixAgent.ps1 -Path . -OutputPath .\powerbi-autonomous-fix-agent.md
.\plugins\powerbi-desktop\scripts\Compare-PowerBILiveRepoModel.ps1 -Path . -OutputPath .\powerbi-live-repo-reconciliation.md
.\plugins\powerbi-desktop\scripts\Test-PowerBIMeasureExpectations.ps1 -Path .
.\plugins\powerbi-desktop\scripts\New-PowerBIPRReleaseComment.ps1 -Path . -OutputPath .\powerbi-pr-comment.md
.\plugins\powerbi-desktop\scripts\New-PowerBIKpiTrustContract.ps1 -Path . -OutputPath .\powerbi-kpi-trust-contract.md
.\plugins\powerbi-desktop\scripts\Invoke-PowerBIAskModel.ps1 -Path . -Question "Which KPIs block release?"
.\plugins\powerbi-desktop\scripts\New-PowerBIFabricReadinessPlan.ps1 -Path . -OutputPath .\powerbi-fabric-readiness.md
.\plugins\powerbi-desktop\scripts\Invoke-PowerBIMaxAIReview.ps1 -Path . -OutputDirectory .\powerbi-max-ai-review

These commands turn review results into Codex-native engineering artifacts: fix loop plans, live/source-control drift, semantic test expectations, PR comments, KPI trust contracts, local model Q&A, and Fabric deployment readiness. They do not sign in to Fabric or publish content.

Run enterprise release engineering

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIReleaseCandidatePack.ps1 -Path . -OutputDirectory .\powerbi-release-candidate-pack -SkipLive
.\plugins\powerbi-desktop\scripts\New-PowerBIServiceScanner.ps1 -Path . -WorkspaceName "Target Workspace" -OutputPath .\powerbi-service-scanner.md
.\plugins\powerbi-desktop\scripts\New-PowerBIModelRiskHeatmap.ps1 -Path . -OutputPath .\powerbi-model-risk-heatmap.md
.\plugins\powerbi-desktop\scripts\Invoke-PowerBISemanticTestRunner.ps1 -Path . -OutputPath .\powerbi-semantic-tests.md
.\plugins\powerbi-desktop\scripts\New-PowerBITomWritePlan.ps1 -Operation AddMeasure -TableName Sales -ObjectName "Average Sales" -Expression "DIVIDE([Total Sales], COUNTROWS('Sales'))"

Use this when the user wants the strongest Power BI/Fabric engineering workflow. The release candidate pack combines unified review, Max AI review, service scanner, semantic tests, model risk heatmap, and PR release comment. Add -IncludeAnalyticalQa when the handoff also needs methodology validation, metric change diagnosis, and an analytical release report. Add -IncludeAdvancedUspQa when the release needs evidence graph, visual-to-measure impact, semantic contract testing, executive trust brief, DAX change risk, freshness/lineage gate, KPI drift watchlist, RLS trust review, UX regression scanning, and migration readiness. Fabric REST and TOM/TMSL writes stay explicit and gated; do not sign in, publish, or mutate a model unless the user explicitly asks and provides the required endpoint/token/write confirmation.

Run analytical release QA

Run:

.\plugins\powerbi-desktop\scripts\Test-PowerBIAnalysisMethodology.ps1 -Path . -OutputPath .\powerbi-analysis-methodology.md
.\plugins\powerbi-desktop\scripts\New-PowerBIMetricChangeDiagnosis.ps1 -Path . -MetricName "Total Sales" -OutputPath .\powerbi-metric-diagnosis.md
.\plugins\powerbi-desktop\scripts\New-PowerBIAnalyticalReleaseReport.ps1 -Path . -OutputPath .\powerbi-analytical-release-report.md
.\plugins\powerbi-desktop\scripts\New-PowerBIReleaseCandidatePack.ps1 -Path . -IncludeAnalyticalQa

Use this when generated review evidence must become stakeholder-ready. Methodology validation checks definitions, grain, denominators, semantic-test coverage, PBIP readiness, and caveats. Metric change diagnosis separates verified movement from likely drivers and missing comparison evidence. The analytical release report turns trust, release-gate, and methodology evidence into a concise Markdown handoff.

Run advanced release USPs

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIEvidenceGraph.ps1 -Path . -OutputPath .\powerbi-evidence-graph.json -Json
.\plugins\powerbi-desktop\scripts\Test-PowerBISemanticContract.ps1 -Path . -OutputPath .\powerbi-semantic-contract.json -Json
.\plugins\powerbi-desktop\scripts\New-PowerBIExecutiveTrustBrief.ps1 -Path . -OutputPath .\powerbi-executive-trust-brief.md
.\plugins\powerbi-desktop\scripts\New-PowerBIReleaseCandidatePack.ps1 -Path . -IncludeAdvancedUspQa

Use this when release confidence depends on traceable evidence, visual impact, semantic contracts, DAX change risk, freshness/lineage, KPI drift monitoring, RLS trust, UX regression risk, and migration readiness.

Run separated portfolio, compliance, and operations QA

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIReleaseCandidatePack.ps1 -Path . -IncludePortfolioGovernanceQa
.\plugins\powerbi-desktop\scripts\New-PowerBIReleaseCandidatePack.ps1 -Path . -IncludeComplianceQa
.\plugins\powerbi-desktop\scripts\New-PowerBIReleaseCandidatePack.ps1 -Path . -IncludeOperationsQa

Use portfolio governance for command-center scoring, cost-to-trust optimization, tenant hygiene, and KPI conflict resolution. Use compliance QA for deployment promotion, certification, accessibility, Power Query contracts, and release evidence signatures. Use operations QA for refresh failure diagnosis, semantic test coverage, and Business KPI SLA monitoring.

Run Fabric live read-only QA

Run:

.\plugins\powerbi-desktop\scripts\Get-PowerBIFabricAccessPlan.ps1 -WorkspaceName "Target Workspace" -AccessTokenPath .\token.txt -Json
.\plugins\powerbi-desktop\scripts\Import-PowerBIFabricWorkspaceSnapshot.ps1 -WorkspaceName "Target Workspace" -AccessTokenPath .\token.txt -OutputDirectory .\fabric-snapshot
.\plugins\powerbi-desktop\scripts\New-PowerBIReleaseCandidatePack.ps1 -Path . -SnapshotDirectory .\fabric-snapshot -IncludeFabricLiveQa -IncludeFabricPortfolioQa -IncludeFabricDeploymentQa -IncludeFabricOperationsQa -IncludeFabricGovernanceQa -IncludeFabricExecutiveQa

Fabric live v1 is read-only. Use token-file auth only, prefer existing snapshots for repeatability, and never publish, promote, refresh, rebind, delete, endorse, or mutate Fabric content from these workflows.

Run business process data quality

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIProcessDataMapping.ps1 -Path . -DataPath .\exports -OutputPath .\process-data-mapping.json -Json
.\plugins\powerbi-desktop\scripts\Invoke-PowerBIBusinessProcessDataQuality.ps1 -Path . -DataPath .\exports -OutputDirectory .\powerbi-business-process-dq -ProcessPack All
.\plugins\powerbi-desktop\scripts\New-PowerBIBusinessProcessDQPack.ps1 -Path . -DataPath .\exports -OutputDirectory .\powerbi-business-process-dq

Use this for Order-to-Cash, Procure-to-Pay, Record-to-Report, Hire-to-Retire, Plan-to-Produce, Forecast-to-Deliver, Service-to-Cash, Issue-to-Resolution, Lead-to-Opportunity, and Quote-to-Order checks. It reads local Power BI metadata and local CSV/JSON exports only; missing mappings return NeedsMapping instead of crashing.

Decide whether a report is release-ready

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBITrustReleaseGate.ps1 -Path . -OutputPath .\powerbi-trust-release-gate.md

Use this for the plugin's differentiating trust workflow. It returns a Go/Warn/No-Go decision based on KPI trust score, unresolved P0 fixes, governance score, Copilot readiness, and low-trust KPI count.

Draft new measures or calculated columns

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIMeasureDraft.ps1 -TableName Sales -MeasureName "Average Sales" -Expression "DIVIDE([Total Sales], COUNTROWS('Sales'))"
.\plugins\powerbi-desktop\scripts\New-PowerBICalculatedColumnDraft.ps1 -TableName Sales -ColumnName "Sales Bucket" -Expression "IF('Sales'[Sales Amount] > 1000, ""High"", ""Standard"")"

These commands create TMDL/PBIP-safe drafts and validation guidance. They do not write directly into binary PBIX/PBIT files. Prefer measures for aggregations and calculated columns only when Power Query or source-system columns are not appropriate.

Draft PBIP pages and visuals

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIReportPageDraft.ps1 -PageName "Executive Overview" -Measures "Total Sales","Sales YoY %"
.\plugins\powerbi-desktop\scripts\Add-PowerBIPBIPReportPage.ps1 -PbipPath .\MyReport -PageName "Executive Overview" -Measures "Total Sales","Sales YoY %" -Apply
.\plugins\powerbi-desktop\scripts\New-PowerBIPBIXCompileWorkflow.ps1 -PbipPath .\MyReport -OutputPbix .\MyReport.pbix

Use this only for PBIP/report JSON workflows. Validate the generated page in Power BI Desktop before saving back to PBIX. If pbi-tools is installed, the compile workflow emits a compile command; otherwise use Power BI Desktop Save As PBIX.

Check model best practices

Run:

.\plugins\powerbi-desktop\scripts\Test-PowerBIModelBestPractices.ps1 -Path . -OutputPath .\powerbi-model-best-practices.md

Use this to evaluate metric ownership, business definitions, PBIP readiness, deterministic DAX, and performance patterns. Trust and best-practice thresholds are configurable in rules/powerbi-trust-rules.json.

Run native tool-parity capabilities

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBINativeToolParityReview.ps1 -Path . -OutputDirectory .\powerbi-native-tool-parity

Use this before relying on external tools. It implements native BPA, model compare/documentation, DAX performance heuristics, report layout checks, theme audit, and PBIP source-control planning from local text/live metadata where available.

Run real feature coverage

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBIRealFeatureReview.ps1 -Path . -OutputDirectory .\powerbi-real-feature-review

Use this for the real Power BI feature layer: visual schema checks, render-readiness, live DAX benchmark timing, live DMV/VertiPaq-style analysis, calculation group drafts, relationship drafts, RLS role drafts, Power Query/M drafts, service integration planning, incremental refresh drafts, aggregation drafts, and schema-aware visual planning.

Apply PBIP-safe drafts

Run:

.\plugins\powerbi-desktop\scripts\Apply-PowerBIPBIPMeasureDraft.ps1 -PbipPath .\MyReport -TableName Sales -MeasureName "Average Sales" -Expression "DIVIDE([Total Sales], COUNTROWS('Sales'))" -Apply
.\plugins\powerbi-desktop\scripts\Apply-PowerBIPBIPPowerQueryDraft.ps1 -PbipPath .\MyReport -QueryName DimDate -SourceKind DateTable -Apply
.\plugins\powerbi-desktop\scripts\Invoke-PowerBIPBIPApplyPlan.ps1 -PbipPath .\MyReport

These scripts write draft artifacts under PBIP folders, maintain a draft manifest where applicable, and include rollback guidance. Validate in Power BI Desktop before saving back to PBIX.

Use optional external Power BI validation tools

Run:

.\plugins\powerbi-desktop\scripts\Get-PowerBIExternalToolInventory.ps1
.\plugins\powerbi-desktop\scripts\Invoke-PowerBIExternalToolsReview.ps1 -Path . -OutputDirectory .\powerbi-external-tools-review

Use this to detect Tabular Editor, DAX Studio, ALM Toolkit, Power BI Helper, Model Documenter, PBI.tips tools, and pbi-tools. The plugin's core capabilities are native; external tools are optional validation paths for engine-specific traces, BPA rule packs, or deployment compare scenarios.

Register as a Power BI External Tool

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIExternalToolRegistration.ps1 -OutputPath .\CodexPowerBIWorkbench.pbitool.json
.\plugins\powerbi-desktop\scripts\Install-PowerBIExternalTool.ps1
.\plugins\powerbi-desktop\scripts\Uninstall-PowerBIExternalTool.ps1

Use -Install only when the user explicitly wants to write into the machine-wide Power BI Desktop External Tools folder. Power BI Desktop should be restarted after installation.

Inspect the open Desktop model

Run:

.\plugins\powerbi-desktop\scripts\Get-PowerBIDesktopLiveConnection.ps1
.\plugins\powerbi-desktop\scripts\Get-PowerBILiveModelSummary.ps1 -OutputPath .\powerbi-live-model-summary.md
.\plugins\powerbi-desktop\scripts\Invoke-PowerBILiveAutoReview.ps1 -OutputDirectory .\powerbi-live-auto-review

This reads the local Analysis Services endpoint created by Power BI Desktop. It can inspect metadata and DMV output from the currently open model when Desktop is running. The live auto-review also creates measure validation, metadata governance, refactor suggestions, prioritized fix backlogs, and DAX fix drafts. Do not use it to modify or publish the report unless the user explicitly asks and a supported write path is available.

Generate an AI sales forecast

Run:

.\plugins\powerbi-desktop\scripts\Invoke-PowerBIAIForecast.ps1 -OutputDirectory .\powerbi-ai-forecast -HorizonMonths 3 -Grain CustomerProduct -Backtest

Use this as a read-only forecast workflow from the open Desktop model. It exports live customer/product monthly data when feasible, falls back to customer-hierarchy/product-line if Desktop cannot materialize the fine grain, blends learned backlog conversion, residual demand, seasonality, budget, roll forecast, and sparse-series shrinkage, then writes detail, summary, top-delta, backtest, and model-quality CSV outputs. Import the output as a separate forecast table before drafting DAX measures.

For innovation beyond a static forecast, route the request to the focused autonomous planning skills:

• powerbi-autonomous-planning-loop: closed-loop observe, forecast, detect, simulate, decide, act, learn, repeat.
• powerbi-goal-seeking-planning: reverse planning from budget, roll, margin, cash, or management targets.
• powerbi-constraint-aware-planning: delivery, capacity, stock, margin, cash, and sales-resource constraints.
• powerbi-revenue-digital-twin: scenario simulation and target-gap closure.
• powerbi-autonomous-forecast-agents: backlog, demand, budget, skeptic, risk, and arbitrator perspectives.
• powerbi-autonomous-exception-management: managed cases for gaps, risks, blockers, and biased planning outputs.
• powerbi-revenue-rescue-mode: operational customer/product actions to close revenue gaps.
• powerbi-forecast-trust-market: human override, confidence, and model-vs-human trust loops.
• powerbi-causal-counterfactual-forecasting: causal drivers and what-if alternatives.
• powerbi-self-healing-forecast-governance: demotion, blocking, and safer-baseline routing when AI quality is poor.
• powerbi-planning-memory: durable run, assumption, action, override, actual, and lesson memory.
• powerbi-planning-readiness-score: autonomous-readiness scoring and blocker identification.
• powerbi-forecast-war-room: executive control tower, action ownership, and gap status.

Customize governance rules

Rules live in rules/powerbi-governance-rules.json. Adjust severities, regex patterns, and thresholds there instead of editing scanner code. Keep custom rules specific enough to avoid noisy findings.

Generate a report theme

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBITheme.ps1 -Name "Executive Analytics" -OutputPath .\executive-theme.json

Use generated themes as a starting point for consistent report styling. Ask before replacing any production theme file.

Generate a metric catalog

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIMetricCatalog.ps1 -Path . -OutputPath .\powerbi-metric-catalog.md

Use this as a semantic contract starter. The generated owner and business definition fields are intentionally TODOs so humans can confirm metric accountability and meaning.

Create a refactoring plan

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIRefactorPlan.ps1 -Path . -OutputPath .\powerbi-refactor-plan.md

Use this to turn scan findings into phased work: Stabilize, Govern, and Polish. Keep it separate from direct file edits until the user confirms the intended changes.

Create a report blueprint

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIReportBlueprint.ps1 -Path . -OutputPath .\powerbi-report-blueprint.md

Use this to convert the metric catalog into a page-level UX plan. It is a design artifact, not a Power BI file writer.

Create an AI prompt pack

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIAIPromptPack.ps1 -Path . -OutputDirectory .\powerbi-ai-pack

Use this when the user wants AI-assisted report improvement. It packages scan results, metrics, dependencies, refactor plan, and report blueprint into context-pack.json plus focused prompts.

Analyze measure impact

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIDependencyGraph.ps1 -Path . -OutputPath .\powerbi-dependency-graph.md
.\plugins\powerbi-desktop\scripts\New-PowerBIDependencyGraph.ps1 -Path . -Mermaid -OutputPath .\powerbi-dependency-graph.mmd

Use this before changing shared measures. Treat hub metrics as higher-risk because dependent measures may inherit changed semantics.

Generate an executive narrative

Run:

.\plugins\powerbi-desktop\scripts\New-PowerBIExecutiveNarrative.ps1 -Path . -OutputPath .\powerbi-executive-narrative.md

Use this for stakeholder-facing summaries. It is deterministic and based on local scan outputs, not a substitute for business-owner sign-off.

Propose changes

• For DAX changes, provide the exact measure expression and target table.
• For Power Query changes, provide the target query and replacement M steps.
• For model changes, describe whether they should be made in Power BI Desktop, Tabular Editor, or source-controlled PBIP files.
• Include a rollback note for report-critical changes.

Validate

When tools are available, prefer:

• Power BI Desktop for opening and visual validation.
• Tabular Editor CLI for Best Practice Analyzer checks.
• DAX Studio or XMLA endpoints for query/performance investigation.
• Source-control diffs for PBIP/TMDL changes.

Script References

• scripts/Test-PowerBIEnvironment.ps1 checks for Power BI Desktop, Tabular Editor, DAX Studio, and common Microsoft Store install locations.
• scripts/New-PowerBIFeatureMaturityMap.ps1 separates implemented, live-read, snapshot-backed, draft/apply, metadata-only, synthetic, and heuristic-simulation capabilities.
• scripts/Get-PowerBIInventory.ps1 inventories PBIX, PBIT, PBIP, TMDL, DAX, Power Query, and model metadata files.
• scripts/New-PowerBIModelSummary.ps1 creates a markdown summary from TMDL, DAX, Power Query, and model.bim files.
• scripts/Invoke-PowerBIInsightScan.ps1 creates a governance and risk report with DAX and Power Query heuristics.
• scripts/New-PowerBITheme.ps1 creates a governed Power BI theme JSON starter.
• scripts/New-PowerBIMetricCatalog.ps1 creates a markdown or JSON metric catalog from DAX/TMDL measures.
• scripts/New-PowerBIRefactorPlan.ps1 converts insight findings into a phased refactoring backlog.
• scripts/New-PowerBIReportBlueprint.ps1 creates a page-level report UX blueprint from the metric catalog.
• scripts/New-PowerBIDependencyGraph.ps1 creates markdown, JSON, or Mermaid dependency graphs for measures.
• scripts/New-PowerBIAIPromptPack.ps1 packages AI-ready context and prompts for review/refactoring/narrative tasks.
• scripts/New-PowerBIExecutiveNarrative.ps1 creates a stakeholder-facing narrative from scan, catalog, and graph outputs.
• scripts/Get-PowerBIPBIPStructure.ps1 scores PBIP/TMDL/report metadata readiness.
• scripts/Invoke-PowerBIAutoReview.ps1 runs the full local AI review package.
• scripts/Get-PowerBIDesktopLiveConnection.ps1 detects the local Desktop model endpoint.
• scripts/Invoke-PowerBILiveDmv.ps1 runs read-only DMV queries against the open Desktop model via ADOMD.
• scripts/Get-PowerBILiveModelSummary.ps1 creates a live model summary from the open Desktop model.
• scripts/New-PowerBILiveMetricCatalog.ps1 creates a live metric catalog from the open Desktop model.
• scripts/New-PowerBILiveDependencyGraph.ps1 creates live measure dependency graphs.
• scripts/Invoke-PowerBILiveInsightScan.ps1 scans the open Desktop model for live governance findings.