Menu
Work with Microsoft Power Automate cloud flows through the power-automate CLI. Use for pull, diff, deploy, create, start, stop, run history, action diagnostics, or replacing a hosted MCP service with local open tooling.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Navigate official Microsoft Power Automate and Power Platform documentation. Use when researching Power Automate capabilities, deciding whether a topic belongs to cloud-flow expressions, Power Fx desktop flows, Dataverse workflow APIs, connector references, limits, errors, or monitoring. Prefer this skill before using generic web search for Power Automate tasks.
Configure Microsoft Power Automate cloud-flow triggers and actions from JSON, connector references, or designer semantics. Use for operation inputs, connector schemas, Office 365 Outlook email actions, SharePoint actions, Dataverse actions, connectionReferences, runAfter, retry policies, secure inputs/outputs, and action payload shape decisions.
Write, validate, and debug Microsoft Power Automate expressions. Use for Workflow Definition Language expressions in cloud flows, dynamic content, runAfter conditions, null-safe property access, variables, trigger/action outputs, date/string/array functions, and for deciding when Power Fx applies only to desktop-flow or other Power Platform formula surfaces.
Authoritative, embedded reference for the Power Automate cloud-flow language (Workflow Definition Language / Logic Apps). Use to confirm whether a function, trigger, or action exists and its exact name or JSON "type", to check function signatures, the workflow-definition schema, runAfter semantics, or operators — before writing or validating any cloud-flow expression or JSON. Stops invented functions and Power Fx bleeding into cloud flows.
Design or modify Microsoft Power Automate cloud-flow architectures using proven workflow patterns. Use when building flows, choosing triggers/actions, structuring branches/scopes, adding SharePoint/Outlook/Dataverse automation, handling retries and errors, or translating n8n-style workflow routines into Power Automate without reinventing the process.
Work with Microsoft Power Platform solutions and solution-scoped environment variables from the command line. Use for listing solutions, inspecting solution components, reading and setting environment variable values (String/Number/Boolean/JSON), handling Secret (Azure Key Vault) variables safely, connection references, and shipping flows through a DEV -> PROD pipeline via the power-automate CLI.
| name | power-automate-api-client |
| description | Work with Microsoft Power Automate cloud flows through the power-automate CLI. Use for pull, diff, deploy, create, start, stop, run history, action diagnostics, or replacing a hosted MCP service with local open tooling. |
Use this skill when changing or diagnosing Power Automate cloud flows from the
command line. The default tool is the power-automate CLI, not a hosted MCP service.
az login session and are never persisted.POWER_AUTOMATE_* environment
variables, then a profiles.json file. See "Target Resolution" below.az login session unless a future task adds
service-principal support.deploy --dry-run before a real deploy.verify.create posts a category-5 cloud flow into the DEV environment.
Always pass --solution (the unmanaged solution, e.g.
<your-solution-unique-name>) so it ships via the DEV->PROD pipeline. Create
in DEV only; never create/deploy on PROD — PROD changes flow exclusively
through the deployment pipeline.power-automate-docs for Microsoft documentation routing.power-automate-run-forensics before changing a flow to fix a runtime
incident.power-automate-expression-syntax and
power-automate-action-configuration before editing definition.power-automate status --profile acme
power-automate pull --profile acme --output ./flow.json
power-automate diff --profile acme flows/<flow>.json --unified
power-automate deploy --profile acme flows/<flow>.json --dry-run
power-automate deploy --profile acme flows/<flow>.json
power-automate create --profile acme flows/<flow>.json --name "<display>" --solution <your-solution-unique-name> [--activate] [--dry-run]
power-automate verify --profile acme flows/<flow>.json
power-automate environments --profile acme
power-automate flows --profile acme
power-automate flows --profile acme --all-environments
power-automate runs --profile acme --top 10
power-automate runs --profile acme --all-environments --top 5
power-automate runs --profile acme --all-environments --json
power-automate run-detail --profile acme <run-id>
power-automate solutions --profile acme [--unmanaged-only]
power-automate solution-components --profile acme --solution <unique-name>
power-automate env-vars --profile acme [--solution <unique-name>]
power-automate env-var-get --profile acme acme_ApiBaseUrl
power-automate env-var-set --profile acme acme_ApiBaseUrl "https://api.example.com" --solution <unique-name> --dry-run
Secret-type environment variables are masked by default (Azure Key
Vault-backed); --reveal-secret shows only the Key Vault reference, never the
secret. env-var-set refuses Secret-type variables. For solution/ALM and
environment-variable work, use the power-automate-solutions skill.
Default discovery stays within the configured profile environment. Use
--all-environments only when the task explicitly needs every environment the
signed-in Azure user can see.
The CLI resolves the non-secret target values in this precedence order:
--environment-id, --flow-id, --dataverse-url,
--entra-tenant-id.POWER_AUTOMATE_ENVIRONMENT_IDPOWER_AUTOMATE_FLOW_IDPOWER_AUTOMATE_DATAVERSE_URLPOWER_AUTOMATE_ENTRA_TENANT_ID~/.config/power-automate-cli/profiles.json, selected
with --profile <profile>:{
"acme": {
"environment_id": "00000000-0000-0000-0000-000000000000",
"flow_id": "00000000-0000-0000-0000-000000000000",
"dataverse_url": "https://org.crm4.dynamics.com",
"entra_tenant_id": "00000000-0000-0000-0000-000000000000"
}
}
dataverse_url and entra_tenant_id are optional. If dataverse_url is
absent, the CLI attempts discovery from the Power Platform environment metadata.
Short-lived OAuth tokens are never stored in any of these locations — they are
sourced exclusively from your current az login session.