OverviewHistoryStatsSecurity
npx skills add ...
Documentation
SKILL.md
npx skills add github/awesome-copilot --skill flowstudio-power-automate-debug
Debug failing Power Automate cloud flows using the FlowStudio MCP server. The Graph API only shows top-level status codes. This skill gives your agent action-level inputs and outputs to find the actual root cause. Load this skill when asked to: debug a flow, investigate a failed run, why is this flow failing, inspect action outputs, find the root cause of a flow error, fix a broken Power Automate flow, diagnose a timeout, trace a DynamicOperationRequestFailure, check connector auth errors, read error details from a run, or troubleshoot expression failures. Requires a FlowStudio MCP subscription — see https://mcp.flowstudio.app
npx skills add github/awesome-copilot --skill flowstudio-power-automate-debug
A step-by-step diagnostic process for investigating failing Power Automate cloud flows through the FlowStudio MCP server.
Real debugging examples: Expression error in child flow | Data entry, not a flow bug | Null value crashes child flow
Prerequisite: A FlowStudio MCP server must be reachable with a valid JWT.
See the flowstudio-power-automate-mcp skill for connection setup.
Subscribe at https://mcp.flowstudio.app
Always call
list_skills/tool_searchfirst to confirm available tool names and parameter schemas. Tool names and parameters may change between server versions. This skill covers response shapes, behavioral notes, and diagnostic patterns — things tool schemas cannot tell you. If this document disagrees withtool_searchor a real API response, the API wins.
CRITICAL:
get_live_flow_run_errortells you which action failed.get_live_flow_run_action_outputstells you why. You must call BOTH. Never stop at the error alone — error codes likeActionFailed,NotSpecified, andInternalServerErrorare generic wrappers. The actual root cause (wrong field, null value, HTTP 500 body, stack trace) is only visible in the action's inputs and outputs.
This is the most important step.
get_live_flow_run_erroronly gives you a generic error code. The actual error detail — HTTP status codes, response bodies, stack traces, null values — lives in the action's runtime inputs and outputs. Always inspect the failing action immediately after identifying it.
Error code from get_live_flow_run_error | What get_live_flow_run_action_outputs reveals |
|---|---|
ActionFailed | Which nested action actually failed and its HTTP response |
NotSpecified | The HTTP status code + response body with the real error |
InternalServerError | The server's error message, stack trace, or API error JSON |
InvalidTemplate | The exact expression that failed and the null/wrong-type value |
BadRequest | The request body that was sent and why the server rejected it |
When actionName refers to an action inside a foreach, the output tool can
return every repetition of that action. Each item may include
repetitionIndexes with the loop name and zero-based itemIndex. Use
iterationIndex to inspect one iteration after you find the suspicious item:
For uncertain connector work, add a Compose_*_Request before the risky action
and a Compose_*_Result after it, with the result action allowed on both
Succeeded and Failed. This gives future debugging a clean payload snapshot
without requiring another deploy. Do not include secrets or long binary payloads
in these bookends.
Find the failing action in the definition. Inspect its inputs expression
to understand what data it expects.
When the failing action's inputs reference upstream actions, inspect those too. Walk backward through the chain until you find the source of the bad data:
⚠️ Output payloads from array-processing actions can be very large. Always slice (e.g.
[:500]) before printing.
Tip: Omit
actionNameto list top-level actions when you're not sure which action produced the bad data. Once you pick an action inside a foreach, passiterationIndexto avoid pulling every repetition into context.
split on null)If the error mentions InvalidTemplate or a function name:
Expression triggerBody()?['fieldName'] returns null → fieldName is wrong.
Inspect the trigger output to see the actual field names:
The error code says InternalServerError or NotSpecified — always inspect
the action outputs to get the actual HTTP status and response body:
Look for ConnectionAuthorizationFailed — the connection owner must match the
service account running the flow. Cannot fix via API; fix in PA designer.
DynamicListValuesUndefinedOrInvalid)Outlook actions like GetEmailsV3 use parameters (mailboxAddress, to, cc,
from) whose dropdown is backed by builtInOperation:AadGraph.GetUsers — which
is broken at the PA listEnum layer and always returns
DynamicListValuesUndefinedOrInvalid. This shows up when an agent rebuilds or
modifies an Outlook action via update_live_flow and tries to resolve a user
through dynamic options. Don't fix it by retrying AadGraph — switch to
shared_office365users.SearchUserV2 instead (returns the same AAD user shape).
Use describe_live_connector to confirm whether the affected parameter exposes
a structured fallback, then call get_live_dynamic_options against
shared_office365users.SearchUserV2 instead of the broken AadGraph operation.
For dynamic field schemas rather than dropdown options, use
get_live_dynamic_properties with the metadata returned by
describe_live_connector.
For expression/data issues:
⚠️
update_live_flowalways returns anerrorkey. A value ofnull(PythonNone) means success.
Use
resubmit_live_flow_runto test ANY flow — not just HTTP triggers.resubmit_live_flow_runreplays a previous run using its original trigger payload. This works for every trigger type: Recurrence, SharePoint "When an item is created", connector webhooks, Button triggers, and HTTP triggers. You do NOT need to ask the user to manually trigger the flow or wait for the next scheduled run.The only case where
resubmitis not available is a brand-new flow that has never run — it has no prior run to replay.
| Scenario | Use | Why |
|---|---|---|
| Testing a fix on any flow | resubmit_live_flow_run | Replays the exact trigger payload that caused the failure — best way to verify |
| Recurrence / scheduled flow | trigger_live_flow (no body) | Runs it now, like the portal's "Run flow" button; resubmit replays a past run's data |
| SharePoint / connector trigger | resubmit_live_flow_run | Cannot be triggered without creating a real SP item |
| HTTP, Button, or PowerApps trigger with custom test payload | trigger_live_flow | When you need to send different data than the original run |
| Brand-new flow, never run | trigger_live_flow | No prior run exists to resubmit |
For flows with a Request trigger (HTTP request, manual Button, or PowerApps),
use trigger_live_flow when you need to send a different payload than the
original run. Pass trigger inputs as body for every kind:
trigger_live_flowhandles AAD-authenticated triggers automatically. Works forRequesttriggers (HTTP request, Button, PowerApps) and for scheduled (Recurrence) flows, which it runs immediately — with nobody, since a scheduled trigger takes no inputs (a body is refused). Automated connector triggers only fire from their source event.Power Automate does not enforce a trigger's
requiredinputs. If you leave one out the run still starts, with that input null, and the result carries awarningnaming the missing keys. Cancel the run and call again with the full body if that matters.
runNameis only returned for Button and PowerApps runs. For HTTP triggers find the run withget_live_flow_runs.Over a browser-extension key, Button and PowerApps triggers run only with an empty body. The tool says so and lists the ways round it: resubmit a past run, default the inputs inside the flow with
coalesce(triggerBody()?['x'], 'value'), or use a standard API key.
| Symptom | First Tool | Then ALWAYS Call | What to Look For |
|---|---|---|---|
| Flow shows as Failed | get_live_flow_run_error | get_live_flow_run_action_outputs on the failing action | HTTP status + response body in outputs |
Error code is generic (ActionFailed, NotSpecified) | — | get_live_flow_run_action_outputs | The outputs.body contains the real error message, stack trace, or API error |
| HTTP action returns 500 | — | get_live_flow_run_action_outputs | outputs.statusCode + outputs.body with server error detail |
| Expression crash | — | get_live_flow_run_action_outputs on prior action | null / wrong-type fields in output body |
| Flow never starts | get_live_flow | — | check properties.state = "Started" |
| Action returns wrong data | get_live_flow_run_action_outputs | — | actual output body vs expected |
| Fix applied but still fails | get_live_flow_runs after resubmit | — | new run status field |
Rule: never diagnose from error codes alone.
get_live_flow_run_erroridentifies the failing action.get_live_flow_run_action_outputsreveals the actual cause. Always call both.
flowstudio-power-automate-mcp — Foundation skill: connection setup, MCP helper, tool discoveryflowstudio-power-automate-build — Build and deploy new flowsresult = mcp("list_live_flows", environmentName=ENV)
# Returns a wrapper object: {mode, flows, totalCount, error}
target = next(f for f in result["flows"] if "My Flow Name" in f["displayName"])
FLOW_ID = target["id"] # plain UUID — use directly as flowName
print(FLOW_ID)runs = mcp("get_live_flow_runs", environmentName=ENV, flowName=FLOW_ID, top=5)
# Returns direct array (newest first):
# [{"name": "08584296068667933411438594643CU15",
# "status": "Failed",
# "startTime": "2026-02-25T06:13:38.6910688Z",
# "endTime": "2026-02-25T06:15:24.1995008Z",
# "triggerName": "manual",
# "error": {"code": "ActionFailed", "message": "An action failed..."}},
# {"name": "...", "status": "Succeeded", "error": null, ...}]
for r in runs:
print(r["name"], r["status"], r["startTime"])
RUN_ID = next(r["name"] for r in runs if r["status"] == "Failed")err = mcp("get_live_flow_run_error",
environmentName=ENV, flowName=FLOW_ID, runName=RUN_ID)
# Returns:
# {
# "runName": "08584296068667933411438594643CU15",
# "failedActions": [
# {"actionName": "Apply_to_each_prepare_workers", "status": "Failed",
# "error": {"code": "ActionFailed", "message": "An action failed..."},
# "startTime": "...", "endTime": "..."},
# {"actionName": "HTTP_find_AD_User_by_Name", "status": "Failed",
# "code": "NotSpecified", "startTime": "...", "endTime": "..."}
# ],
# "allActions": [
# {"actionName": "Apply_to_each", "status": "Skipped"},
# {"actionName": "Compose_WeekEnd", "status": "Succeeded"},
# ...
# ]
# }
# failedActions is ordered outer-to-inner. The ROOT cause is the LAST entry:
root = err["failedActions"][-1]
print(f"Root action: {root['actionName']} → code: {root.get('code')}")
# allActions shows every action's status — useful for spotting what was Skipped
# See common-errors.md to decode the error code.# Get the root failing action's full inputs and outputs
root_action = err["failedActions"][-1]["actionName"]
detail = mcp("get_live_flow_run_action_outputs",
environmentName=ENV,
flowName=FLOW_ID,
runName=RUN_ID,
actionName=root_action)
if len(detail) > 1:
print(f"{root_action} returned {len(detail)} repetitions; inspect iteration indexes")
out = detail[0] if detail else {}
print(f"Action: {out.get('actionName')}")
print(f"Status: {out.get('status')}")
# For HTTP actions, the real error is in outputs.body
if isinstance(out.get("outputs"), dict):
status_code = out["outputs"].get("statusCode")
body = out["outputs"].get("body", {})
print(f"HTTP {status_code}")
print(json.dumps(body, indent=2)[:500])
# Error bodies are often nested JSON strings — parse them
if isinstance(body, dict) and "error" in body:
err_detail = body["error"]
if isinstance(err_detail, str):
err_detail = json.loads(err_detail)
print(f"Error: {err_detail.get('message', err_detail)}")
# For expression errors, the error is in the error field
if out.get("error"):
print(f"Error: {out['error']}")
# Also check inputs — they show what expression/URL/body was used
if out.get("inputs"):
print(f"Inputs: {json.dumps(out['inputs'], indent=2)[:500]}")all_reps = mcp("get_live_flow_run_action_outputs",
environmentName=ENV,
flowName=FLOW_ID,
runName=RUN_ID,
actionName=root_action)
for rep in all_reps[:10]:
print(rep.get("repetitionIndexes"), rep.get("status"), rep.get("error"))
one_rep = mcp("get_live_flow_run_action_outputs",
environmentName=ENV,
flowName=FLOW_ID,
runName=RUN_ID,
actionName=root_action,
iterationIndex=3)Error code: "InternalServerError" ← this tells you nothing
Action outputs reveal:
HTTP 500
body: {"error": "Cannot read properties of undefined (reading 'toLowerCase')
at getClientParamsFromConnectionString (storage.js:20)"}
← THIS tells you the Azure Function crashed because a connection string is undefinedError code: "BadRequest" ← generic
Action outputs reveal:
inputs: "body('HTTP_GetTokenFromStore')?['token']?['access_token']"
outputs: "" ← empty string, the path resolved to null
← THIS tells you the response shape changed — token is at body.access_token, not body.token.access_tokendefn = mcp("get_live_flow", environmentName=ENV, flowName=FLOW_ID)
actions = defn["properties"]["definition"]["actions"]
print(list(actions.keys()))# Inspect multiple actions leading up to the failure
for action_name in [root_action, "Compose_WeekEnd", "HTTP_Get_Data"]:
result = mcp("get_live_flow_run_action_outputs",
environmentName=ENV,
flowName=FLOW_ID,
runName=RUN_ID,
actionName=action_name)
out = result[0] if result else {}
print(f"\n--- {action_name} ({out.get('status')}) ---")
print(f"Inputs: {json.dumps(out.get('inputs', ''), indent=2)[:300]}")
print(f"Outputs: {json.dumps(out.get('outputs', ''), indent=2)[:300]}")# Example: action uses split(item()?['Name'], ' ')
# → null Name in the source data
result = mcp("get_live_flow_run_action_outputs", ..., actionName="Compose_Names")
if not result:
print("No outputs returned for Compose_Names")
names = []
else:
names = result[0].get("outputs", {}).get("body") or []
nulls = [x for x in names if x.get("Name") is None]
print(f"{len(nulls)} records with null Name")result = mcp("get_live_flow_run_action_outputs", ..., actionName="<trigger-action-name>")
print(json.dumps(result[0].get("outputs"), indent=2)[:500])result = mcp("get_live_flow_run_action_outputs", ..., actionName="HTTP_Get_Data")
out = result[0]
print(f"HTTP {out['outputs']['statusCode']}")
print(json.dumps(out['outputs']['body'], indent=2)[:500])defn = mcp("get_live_flow", environmentName=ENV, flowName=FLOW_ID)
acts = defn["properties"]["definition"]["actions"]
# Example: fix split on potentially-null Name
acts["Compose_Names"]["inputs"] = \
"@coalesce(item()?['Name'], 'Unknown')"
conn_refs = defn["properties"]["connectionReferences"]
result = mcp("update_live_flow",
environmentName=ENV,
flowName=FLOW_ID,
definition=defn["properties"]["definition"],
connectionReferences=conn_refs)
print(result.get("error")) # None = success# Resubmit the failed run — works for ANY trigger type
resubmit = mcp("resubmit_live_flow_run",
environmentName=ENV, flowName=FLOW_ID, runName=RUN_ID)
print(resubmit) # {"resubmitted": true, "triggerName": "..."}
# Wait ~30 s then check
import time; time.sleep(30)
new_runs = mcp("get_live_flow_runs", environmentName=ENV, flowName=FLOW_ID, top=3)
print(new_runs[0]["status"]) # Succeeded = done# First inspect what the trigger expects — read directly from the flow definition
defn = mcp("get_live_flow", environmentName=ENV, flowName=FLOW_ID)
triggers = defn["properties"]["definition"]["triggers"]
manual = next(iter(triggers.values())) # usually the only trigger on HTTP flows
request_schema = manual.get("inputs", {}).get("schema")
print("Expected body schema:", request_schema)
# Response schemas live on Response action(s) in the actions block
for name, act in defn["properties"]["definition"]["actions"].items():
if act.get("type") == "Response":
print(f"Response {name}:", act.get("inputs", {}).get("schema"))
# Trigger with a test payload
result = mcp("trigger_live_flow",
environmentName=ENV,
flowName=FLOW_ID,
body={"name": "Test User", "value": 42})
print(f"Status: {result['responseStatus']}, Body: {result.get('responseBody')}")
print(f"Kind: {result['triggerKind']}, via: {result['invocation']}, run: {result.get('runName')}")
if result.get("warning"):
print(result["warning"]) # required trigger inputs you left out