flowstudio-power-automate-mcp

Operate Power Automate cloud flows programmatically through a FlowStudio MCP server.

1|1|Updated Jan 29, 2026
One-click install
npx skills add https://github.com/ultraviollettnympho/transit-ticket --skill flowstudio-power-automate-mcp-ultraviollettnympho
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: flowstudio-power-automate-mcp
Source: https://github.com/ultraviollettnympho/transit-ticket/tree/main/.github/skills/flowstudio-power-automate-mcp
Command: npx skills add https://github.com/ultraviollettnympho/transit-ticket --skill flowstudio-power-automate-mcp-ultraviollettnympho

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Managing Microsoft Power Automate flows normally requires clicking through the web portal to list flows, read definitions, debug failed runs, and resubmit executions. This Skill lets an AI agent perform all of those operations programmatically through the FlowStudio MCP server, with no browser or manual UI steps. ## Core Features & Use Cases - Flow Discovery & Inspection: List flows and environments, fetch complete flow definitions including triggers, actions, and connection references. - Run Monitoring & Debugging: Retrieve run history, get structured per-action error breakdowns, and inspect action inputs/outputs to find root causes of failures. - Run Control & Deployment: Resubmit failed runs, cancel running executions, trigger HTTP flows, and create or patch flow definitions via update_live_flow. - Use Case: A flow fails overnight. Ask the agent to find the failed run, identify the root failing action, inspect the upstream action's output for null values, patch the expression with a coalesce fix, redeploy, and resubmit the run to verify success. ## Quick Start Ask the agent to list all Power Automate flows in your environment and show the status of the most recent runs for a specific flow.

Frequently Asked Questions about flowstudio-power-automate-mcp

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I list and debug Power Automate flows without the web portal?

Call the FlowStudio MCP server with tools like list_live_flows, get_live_flow_runs, and get_live_flow_run_error. These return flow lists, run history, and per-action failure details directly from the Power Automate API, with the last entry in failedActions being the root cause.

How do I resubmit a failed Power Automate flow run programmatically?

Use the resubmit_live_flow_run tool with the environmentName, flowName, and runName of the failed run. The run ID comes from the name field in get_live_flow_runs results, and the flow re-executes with its original trigger payload.

What authentication does the FlowStudio MCP server require?

The server requires an x-api-key header containing a plain JWT token from mcp.flowstudio.app, not an Authorization Bearer header. You also need the Power Platform environment name, typically in the format Default-<tenant-guid>, discoverable via list_live_environments.

Can I update a Power Automate flow definition via API?

Yes, update_live_flow patches an existing definition or creates a new flow when flowName is omitted. You must pass the definition, connectionReferences, and a description, and the response error field is null on success.

Why does trigger_live_flow fail for my scheduled flow?

trigger_live_flow only works for HTTP Request triggers and returns an error for Recurrence or connector-based triggers. For non-HTTP flows, use resubmit_live_flow_run on an existing run instead.

When should I not cancel a running Power Automate flow?

Do not cancel a run showing Running status when it is waiting for an adaptive card response in Teams, since that status is normal while paused for human input. Cancelling discards the pending card and its response.