@uipath/skills 1.195.0

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/index-out-of-bounds.md

@@ -0,0 +1,49 @@
+---
+confidence: low
+---
+
+# Index Outside Bounds of Array (502)
+
+## Context
+
+What this looks like:
+- HTTP 502 surfaced by Maestro with `System.IndexOutOfRangeException`
+- Error message: `Index was outside the bounds of the array.`
+- No actionable detail in `errorDetails` beyond the stack trace
+
+What can cause it:
+- Server-side bug — engine accessed an array slot that doesn't exist
+- Incorrect Orchestrator URL (trailing slash, wrong tenant path) causing client init to read past expected segments
+- Culture data initialization failure on the host
+- Re-entry into an already-executed parallel multi-instance subprocess — historical case where a follow-up workflow updated files and tried to re-run prior steps
+
+What to look for:
+- Stack trace in `errorDetails` — what component is throwing
+- Whether the workflow loops back into a parallel multi-instance subprocess
+- Whether the Orchestrator URL in the tenant settings is well-formed
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json` — preserve the full stack trace
+2. Walk element executions: `uip maestro <type> instance element-executions <instance-id> -f <folder-key> --output json` — note the element that faulted
+3. Check whether the workflow attempted to re-enter a multi-instance subprocess that already completed
+4. If reproducible, gather: Orchestrator URL config, exact `.bpmn`, sample inputs
+
+## Resolution
+
+- **If re-entry into completed multi-instance subprocess:** redesign so the follow-up path uses a fresh subprocess instead of re-entering the original
+- **If bad Orchestrator URL:** correct the tenant URL (no trailing slash, correct tenant path)
+- **If stack trace points at culture data:** escalate to platform/infra — server-side fix needed
+- **If no actionable root cause from the stack trace:** open a UiPath support ticket with the full stack trace, the `.bpmn`, the instance ID, and a minimal repro
+
+## Notes
+
+- This error is **Not Troubleshootable** to a user-actionable fix from PIMS API alone — it almost always indicates a platform bug. Capture the stack trace and escalate
+
+## References
+
+- [Docs: Orchestrator Exceptions](https://docs.uipath.com/orchestrator/standalone/2024.10/installation-guide/orchestrator-exceptions)
+- [Forum: Index was outside the bounds](https://forum.uipath.com/t/could-not-create-orchestrator-client-exception-index-was-outside-the-bounds-of-the-array/799616)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/input-schema-mismatch.md

@@ -0,0 +1,47 @@
+---
+confidence: high
+---
+
+# Input Does Not Conform to Schema (400)
+
+## Context
+
+What this looks like:
+- HTTP 400 at workflow/agent start
+- Error message: `Input does not conform to schema` or `Agent.InputArgumentsSchema` error
+- Variants: `"Value is 'null' but should be 'object'"` when an optional file/attachment is null
+- Used to work, suddenly fails after a new version was published
+
+What can cause it:
+- Provided inputs do not match the JSON schema declared on the process/agent: wrong types, missing required fields, extra fields, or unsupported types
+- For queue transactions: `DateTime` must be passed as string; arrays are not supported
+- Optional file/attachment parameters passed as `null` — schema rejects `null` where it expects an `object`
+- Newly published version added a required parameter (e.g., `modelName` on GenAI activities) that older callers don't pass
+- Stale agent metadata from a previous publish — Maestro still references the old `InputArgumentsSchema`
+
+What to look for:
+- Compare the actual payload (Variables API) against the agent's/process's declared input schema (`inputDefinitions`)
+- Confirm whether the failure started after a new package version was deployed
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json`
+2. Pull the actual inputs and the schema: `uip maestro <type> instance variables <instance-id> -f <folder-key> --output json` — compare values vs `inputDefinitions`
+3. Identify the failing parameter from `errorDetails` — the schema validator names the offending field/type
+4. Check the package version that introduced the breaking change — list releases and diff schemas
+
+## Resolution
+
+- **If type/field mismatch:** correct the payload to match the schema exactly (types, required fields, no extras)
+- **If null optional file/attachment:** either omit the field entirely or pass a valid empty-shaped object the schema accepts
+- **If queue payload:** convert `DateTime` to string; flatten arrays
+- **If new required parameter added:** make it optional with a default value, or update all callers
+- **If stale agent metadata:** delete the deployed agent and let the next solution publish re-create it (last-resort workaround)
+
+## References
+
+- [Docs: About Input and Output Arguments](https://docs.uipath.com/orchestrator/standalone/2023.4/user-guide/about-input-and-output-arguments)
+- [Docs: About Queues and Transactions](https://docs.uipath.com/orchestrator/standalone/2024.10/user-guide/about-queues-and-transactions)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/insufficient-funds.md

@@ -0,0 +1,44 @@
+---
+confidence: high
+---
+
+# Insufficient Funds / Agent Units (400)
+
+## Context
+
+What this looks like:
+- HTTP 400 from an agent or GenAI activity
+- Error message: `Insufficient funds: Your account doesn't have enough credits for execution. Please add funds or contact your administrator for assistance.`
+
+What can cause it:
+- Tenant has exhausted its Agent Units (AU) allocation — agents and GenAI activities consume AU on every LLM call
+- Community Edition free-tier credits exhausted (reset monthly)
+- Enterprise subscription on a Flex license without AU allocated
+- Known bug (~early 2026): Maestro debug incorrectly required AU even though standalone Agent debug did not; impacted customers without AU even when they expected debug to be free
+
+What to look for:
+- Whether the failing element is an Agent or GenAI activity — non-agent processes do not consume AU
+- Tenant AU balance — exhausted vs not allocated
+- Debug vs deployed — historical bug only affected debug mode
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json` — error message is self-explanatory
+2. Confirm the element type: agent/GenAI vs other
+3. Check AU balance: Orchestrator UI → **Admin > Tenant > Licenses > Consumption > Agent Units**
+4. Verify debug-mode behaviour against the latest Maestro release notes — the debug-consumes-AU bug had a planned fix in late Feb / mid-March 2026
+
+## Resolution
+
+- **If AU exhausted on Enterprise:** purchase / allocate more via **Admin > Organization > Subscriptions** or contact the UiPath account manager
+- **If Community / Flex with no AU:** request promo units or upgrade to a plan that includes AU
+- **If only failing in debug:** confirm the Maestro deploy includes the fix for the debug-mode AU bug; if not, request rollout or escalate to the Maestro team
+- **If the process should not need AU at all:** verify no Agent/GenAI activities are accidentally present in the path being executed
+
+## References
+
+- [Docs: Purchasing a Plan](https://docs.uipath.com/automation-cloud/automation-cloud/latest/admin-guide/purchasing-a-plan)
+- [Forum: Insufficient funds](https://forum.uipath.com/t/agent-baseerror-insufficient-funds-your-account-doesnt-have-enough-credits-for-execution/5726606)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/integration-service-400.md

@@ -0,0 +1,51 @@
+---
+confidence: medium
+---
+
+# Integration Services 400 — Bad Request to Connector
+
+## Context
+
+What this looks like:
+- HTTP 400 from an Integration Service call
+- Error message: `Request to Integration Services failed with status code '400'`
+- Often paired with a connector-side error body (multipart parse failure, schema mismatch, etc.)
+
+What can cause it:
+- Malformed or invalid request body for the connector action
+- Connector update changed expected `Content-Type` (e.g., Outlook moved from `application/json` to `multipart/form-data` for attachments) breaking already-published activities — historical bug `ENGCE-46376`
+- A previously-bound variable was removed but the connector activity still references it, sending an empty string `""` instead of `null` (Office 365 Send Email historical issue, `PLT-77820`)
+- Mismatched connector input schema after a connector version bump
+- Bad connection state (token, scopes) — connection should be re-provisioned
+
+What to look for:
+- Connector name / action from `errorDetails` and element executions
+- Whether the failure started after a connector update or Maestro publish
+- Whether the same action works against a freshly-recreated connector activity
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json` — capture upstream connector error body
+2. Identify the failing connector activity and connection: `uip maestro <type> instance element-executions <instance-id> -f <folder-key> --output json`
+3. Pull the activity inputs: `uip maestro <type> instance variables <instance-id> -f <folder-key> --parent-element-id <element-id> --output json` — look for empty strings where nulls are expected, missing required fields, or wrong types
+4. Check the connection's health in Orchestrator UI → **Integration Service > Connections** — re-authenticate if expired
+5. Look up the connector version and compare against the version when the workflow was last published
+
+## Resolution
+
+- **If empty string from removed variable:** delete and re-add the connector step so it stops referencing the orphaned variable
+- **If `Content-Type` / multipart mismatch after connector update:** delete and re-add the activity to pick up the new schema
+- **If schema bump:** update the inputs to match the new schema or pin to the prior connector version where supported
+- **If connection in bad state:** re-provision the connection, then re-publish or rebind the activity
+- **If genuinely malformed payload:** validate the input JSON against the connector spec
+
+## Notes
+
+- This error is **Not Troubleshootable** from PIMS API alone — incident lacks the actual HTTP request that went to IS and the connector configuration details. Plan to log a connector telemetry ask if this becomes common in your investigation
+
+## References
+
+- [Docs: Connections Troubleshooting](https://docs.uipath.com/integration-service/automation-cloud/latest/user-guide/connections-troubleshooting)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/integration-service-404.md

@@ -0,0 +1,54 @@
+---
+confidence: medium
+---
+
+# Integration Service Failure (404)
+
+## Context
+
+What this looks like:
+- HTTP 404 from an Integration Service call
+- Error message: `Request to Integration Services failed with status code '404'`
+- Sometimes seen with a related text like `Invalid Element Instance Id provided.` or `Missing instance` (Data Fabric writes)
+
+What can cause it:
+- Connector or action was deleted/renamed since publish
+- Integration Service connection expired or was removed
+- Tenant doesn't have Integration Service enabled (**Admin > Tenant > Services**)
+- Webhook config references a non-existent object on the external side
+- Solution cross-imported between environments (staging → alpha) where the folder key embedded in the solution doesn't exist in the destination — surfaces as 404/403
+- For Data Fabric writes: missing instance, folder, or tenant permission on the target
+
+What to look for:
+- The connector name and action attempted — Variables API + element executions
+- Whether the failure started right after a solution import or connector rename
+- Connection state in the IS UI
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json`
+2. Identify the failing connector activity: `uip maestro <type> instance element-executions <instance-id> -f <folder-key> --output json`
+3. Check the IS connection: Orchestrator UI → **Integration Service > Connections** — verify exists, not expired
+4. Confirm IS is enabled on the tenant
+5. For Data Fabric writes, verify the target entity/instance exists in the destination folder and the robot has permission
+
+## Resolution
+
+- **If connection expired:** re-authenticate; re-publish the process if needed
+- **If connector deleted/renamed:** re-create the activity against the current connector and re-publish
+- **If IS not enabled on the tenant:** enable via **Admin > Tenant > Services**
+- **If webhook config stale:** regenerate the webhook URL and update the external system
+- **If solution cross-imported with stale keys:** update bindings to the destination folder keys and re-deploy
+- **If Data Fabric target missing:** create the entity/instance in the destination folder, or fix the binding
+
+## Notes
+
+- This error is **Not Troubleshootable** from PIMS API alone — incident lacks the connector name, connection ID, and action attempted. Expect to ask the user for the failing element name; planned enrichment in incident data would move this to "Full"
+
+## References
+
+- [Forum: IS resource not found in Community Cloud](https://forum.uipath.com/t/persistent-integration-services-resource-not-found-error-in-community-cloud/5712850)
+- [Docs: Connections Troubleshooting](https://docs.uipath.com/integration-service/automation-cloud/latest/user-guide/connections-troubleshooting)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/job-operation-timeout.md

@@ -0,0 +1,44 @@
+---
+confidence: medium
+---
+
+# Job Operation Timeout (502)
+
+## Context
+
+What this looks like:
+- HTTP 502 with error message `Job Operation Timeout`
+- Job appears completed in Orchestrator logs but Maestro reports timeout
+- Instance stuck several minutes past the underlying job's actual end time
+
+What can cause it:
+- A gateway/proxy (Azure Application Gateway, Cloudflare, nginx) times out before Orchestrator returns. Typical with synchronous call mode where the HTTP connection is held until the job completes
+- Many concurrent action-center tasks each spawning a job — resource contention pushes individual call latency past the gateway threshold
+- Serverless/coded RPA timing out at the serverless layer ("automation cancelled because it reached the time limit") — root cause is on the robot side, not Maestro (tracked in past Slack threads as `MST-6379`)
+
+What to look for:
+- Whether the underlying child job is actually `Successful` in Orchestrator while Maestro reports timeout — indicates gateway timeout, not runtime failure
+- Whether the same workflow times out on subsequent runs — points at sustained gateway/proxy limits
+- Concurrent task volume on the folder
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json`
+2. Pull element executions and capture start/complete times of the timing-out activity: `uip maestro <type> instance element-executions <instance-id> -f <folder-key> --output json`
+3. Locate the child Orchestrator job ID from the incident `errorDetails` and check its final state: `uip or jobs get <child-job-key> --output json`
+4. If the child job is `Successful` but Maestro 502'd, the gateway timed out before Orchestrator's response landed
+
+## Resolution
+
+- **If using synchronous call mode:** switch to **asynchronous** call mode — Maestro polls the status URI instead of holding a connection
+- **If behind a reverse proxy:** increase the request timeout on the backend (Application Gateway / nginx / Cloudflare)
+- **If serverless/coded RPA timeout:** check the serverless robot logs, raise the time limit on the coded automation, or split the workflow
+- **If many concurrent jobs:** stagger the triggers or add a multi-instance marker with a smaller batch instead of one job per task
+
+## References
+
+- [Docs: Call modes explained](https://docs.uipath.com/orchestrator/automation-cloud/latest/user-guide/call-modes-explained)
+- [Docs: Frequently Encountered Orchestrator Errors](https://docs.uipath.com/orchestrator/docs/frequently-encountered-orchestrator-errors)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/js-runtime-discrepancy.md

@@ -0,0 +1,26 @@
+---
+confidence: high
+---
+
+# JS Runtime Discrepancy
+
+## Context
+
+What this looks like:
+- JavaScript expression passes in JS Editor but fails at runtime
+- "btoa is not defined" or other runtime JS errors
+- Browser-specific APIs (TextEncoder, atob, btoa) fail at runtime
+
+What can cause it:
+- The design-time JS Editor runs in the browser (supports btoa, TextEncoder, etc.) but runtime uses Jint (.NET JS interpreter) which lacks browser-specific APIs
+
+## Investigation
+
+1. Identify which JS API the expression uses
+2. Check if that API is a browser-specific API not supported by Jint
+
+## Resolution
+
+- Use Jint-compatible alternatives. For base64 encoding use: `new Uint8Array(Array.from("test", c => c.charCodeAt(0))).toBase64()`
+- Reference the Jint GitHub repo for supported features
+- Validate the expression works in both the JS Editor and at runtime

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/loop-detected.md

@@ -0,0 +1,44 @@
+---
+confidence: high
+---
+
+# Element Execution Loop Detected (400009)
+
+## Context
+
+What this looks like:
+- HTTP 400, Maestro error code `400009` (`ElementExecutionLoopDetected`)
+- Error message: `Possible loop detected: element 'X' has been executed more than 100 times. Failing the instance to prevent infinite loop.`
+- Process was making progress, then engine kills the instance at iteration 101
+
+What can cause it:
+- Element exceeded `MaxElementExecutionCount` (default: 100)
+- A control-loop exit condition never evaluates `true` (counter not incrementing, wrong comparison, variable not updated inside loop body)
+- Legitimate large-batch workflow (e.g., 2000 queue items processed via gateway loop instead of marker)
+
+What to look for:
+- Element name and iteration count from `errorDetails`
+- The exit condition expression — does it actually reference the variable that the loop body updates?
+- Whether the workflow uses an explicit gateway loop or a multi-instance marker
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json` — captures element ID and iteration count
+2. Walk element executions to see what each iteration produced: `uip maestro <type> instance element-executions <instance-id> -f <folder-key> --output json`
+3. Pull the variables snapshot near the loop element: `uip maestro <type> instance variables <instance-id> -f <folder-key> --parent-element-id <element-id> --output json` — look for the loop counter / exit condition variable
+4. Read the loop condition from the BPMN: `uip maestro <type> instance asset <instance-id> -f <folder-key> --output json`
+
+## Resolution
+
+- **If counter not incrementing or wrong operator:** fix the loop body so the exit condition can eventually evaluate `true`
+- **If legitimate large-batch processing:** redesign with a multi-instance marker (parallel or sequential) instead of a gateway loop — markers avoid loop detection entirely because there is no explicit loop element in the BPMN
+- **If process genuinely needs >100 iterations of one element:** request the server-side `MaxElementExecutionCount` be raised via an environment-level feature flag (platform team)
+- **If variable not updated inside loop body:** ensure the assignment happens before the gateway re-evaluates
+
+## References
+
+- [Docs: Looping](https://docs.uipath.com/maestro/automation-cloud/latest/user-guide/repetition)
+- [Docs: BPMN Patterns and Practices](https://docs.uipath.com/maestro/automation-cloud/latest/user-guide/common-modeling-patterns)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/maestro-service-disabled.md

@@ -0,0 +1,26 @@
+---
+confidence: high
+---
+
+# Maestro Service Disabled
+
+## Context
+
+What this looks like:
+- Error opening any Maestro workflow
+- Designer pane is blank in Studio Web
+- Studio Web shows errors after a license change
+
+What can cause it:
+- License entitlement revoked by OMS silently disables the Maestro service on the tenant
+
+## Investigation
+
+1. Check if the Maestro service is enabled on the tenant admin dashboard
+2. Check if the user's license includes Maestro access
+
+## Resolution
+
+- Re-enable the Maestro service from the Admin dashboard on the affected tenant
+- Verify license options are correctly assigned
+- Confirm all Maestro workflows can be opened after re-enabling

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/marker-input-null.md

@@ -0,0 +1,45 @@
+---
+confidence: high
+---
+
+# Marker Input Collection Is Null (400007)
+
+## Context
+
+What this looks like:
+- HTTP 400, Maestro error code `400007` (`BpmnMarkerInputNullError`)
+- Error message: `Input collection for the marker element must not be null`
+- Sometimes seen alongside the related evaluation error `400008` (`BpmnMarkerInputEvaluationFailure`) — see [marker-invalid-cast](marker-invalid-cast.md) for the JS array cast bug
+
+What can cause it:
+- The variable referenced in the multi-instance marker's "Items" property is `null` at runtime (never assigned, or upstream task didn't produce the list)
+- A `NullReferenceException` in `BpmnExecution.cs` where a marker input was assumed non-null but the underlying success value was null (known internal bug; tracked historically — fix replaces `!` null-forgiving with `?? throw`)
+- User Task multi-instance "items" passed a malformed JSON array (e.g., a stringified array instead of a real array)
+
+What to look for:
+- The marker's `elementId` from the incident
+- Variables API: confirm the collection variable is actually `null` (vs empty array, which behaves differently)
+- Upstream element that should have produced the collection
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json`
+2. Pull variables scoped to the marker element: `uip maestro <type> instance variables <instance-id> -f <folder-key> --parent-element-id <marker-element-id> --output json` — verify the "Items" variable is null
+3. Walk element executions to find the upstream task that was supposed to produce the collection: `uip maestro <type> instance element-executions <instance-id> -f <folder-key> --output json`
+4. Inspect the marker's "Items" expression in the BPMN: `uip maestro <type> instance asset <instance-id> -f <folder-key> --output json`
+
+## Resolution
+
+- **If upstream task produced nothing:** fix the upstream task or add an exclusive gateway that skips the marker when the collection is null/empty
+- **If variable was renamed / unbound:** re-select the collection variable in the marker's "Items" property
+- **If malformed JSON array on User Task:** ensure callers construct a real JSON array, not a stringified one
+- **If you suspect the internal null-forgiving bug:** request the Maestro engine team confirm the `BpmnExecution.cs` fix is deployed in your region; in the meantime ensure the upstream activity never returns a null wrapper
+- **If using JS expression and seeing `InvalidCastException`:** switch to C# expressions and see [marker-invalid-cast](marker-invalid-cast.md)
+
+## References
+
+- [Docs: Multi-instance markers](https://docs.uipath.com/maestro/automation-cloud/latest/user-guide/markers-implementation)
+- [Docs: Markers](https://docs.uipath.com/maestro/automation-cloud/latest/user-guide/markers)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/marker-invalid-cast.md

@@ -0,0 +1,38 @@
+---
+confidence: high
+---
+
+# Multi-Instance Marker InvalidCastException (400008)
+
+## Context
+
+What this looks like:
+- HTTP 400, Maestro error code `400008` (`BpmnMarkerInputEvaluationFailure`)
+- Error message: `Failed to evaluate the input collection variable for the marker element`
+- Inner exception: `InvalidCastException: System.Object[] to ExpressionList`
+- Sibling: see [marker-input-null](marker-input-null.md) when the collection is null (error `400007`)
+
+What can cause it:
+- Bug in `JavaScriptEvaluator.cs` — JS expression returns `Object[]` which cannot be cast to the internal `ExpressionList` type the BPMN engine expects on a multi-instance marker
+
+What to look for:
+- Whether the marker's "Items" expression uses `=js:` (JavaScript) — that's the trigger
+- Whether the same logic written as C# works
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json` — confirm `InvalidCastException` and `ExpressionList`
+2. Check the marker's "Items" expression language in the BPMN: `uip maestro <type> instance asset <instance-id> -f <folder-key> --output json`
+
+## Resolution
+
+- Switch the marker's "Items" expression from JavaScript to a C# expression that returns a typed list/array
+- Re-publish and re-run; the parallel marker should iterate all items
+- If switching to C# is not feasible, work around by materializing the JS array into a typed collection variable in a preceding Script task
+
+## References
+
+- [Docs: Multi-instance markers](https://docs.uipath.com/maestro/automation-cloud/latest/user-guide/markers-implementation)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/missing-required-parameter.md

@@ -0,0 +1,48 @@
+---
+confidence: high
+---
+
+# Missing Value for Required Parameter (400)
+
+## Context
+
+What this looks like:
+- HTTP 400 at activity execution
+- Error message: `Missing value for required parameter 'id'` (or `'worksheetID'`, `'modelName'`, etc.)
+- Often appears immediately after an Integration Service connector update or a Maestro release
+
+What can cause it:
+- An API call has a URL template `{id}` placeholder that wasn't substituted
+- A required input argument named in the schema was not included in the payload
+- Maestro UI is missing a property that exists in Studio/Studio Web (UI bug) — user can't set the value because the field isn't shown
+- A connector or activity update added a new required property (e.g., GenAI `modelName`), breaking already-published workflows
+- Input argument JSON exceeded the 10,000-character limit and got truncated, dropping the parameter
+
+What to look for:
+- Match the missing parameter name from the error to the activity's expected inputs
+- Compare actual inputs (Variables API) vs `inputDefinitions`
+- Whether the workflow worked before a recent connector/Maestro release
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json` — error message names the missing parameter
+2. Pull the activity inputs: `uip maestro <type> instance variables <instance-id> -f <folder-key> --output json`
+3. Compare to the activity's `inputDefinitions` in the BPMN/`.bpmn` asset: `uip maestro <type> instance asset <instance-id> -f <folder-key> --output json`
+4. Check the Maestro UI for the failing activity — is the missing field even visible? If not, suspect a Maestro frontend bug
+5. Verify the input JSON size — over 10,000 characters indicates truncation
+
+## Resolution
+
+- **If parameter genuinely missing:** add it to the activity's input mapping
+- **If URL `{id}` placeholder:** ensure the ID value is substituted, not the literal placeholder
+- **If Maestro UI doesn't expose the field:** workaround — manually edit the `.bpmn` to add the property, or wait for the FE fix
+- **If new required property from connector update:** make it optional with a default (connector owner) or update all callers to pass the value; check that pre-hook defaults are deployed in the target region
+- **If JSON truncation:** reduce input size or split into multiple smaller arguments
+
+## References
+
+- [Docs: Building API Requests](https://docs.uipath.com/orchestrator/automation-cloud/latest/api-guide/building-api-requests)
+- [Forum: StartJobs API call error code 400](https://forum.uipath.com/t/startjobs-api-call-error-code-400/469915)

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/multi-instance-parallel.md

@@ -0,0 +1,37 @@
+---
+confidence: medium
+---
+
+# Multi-Instance Parallel Marker Issues
+
+## Context
+
+What this looks like:
+- "Failed to evaluate the input collection variable for the marker element"
+- Parallel steps failing
+- Questions about how to loop over items
+
+What can cause it:
+- JS expression InvalidCastException (see [marker-invalid-cast](marker-invalid-cast.md) for the specific bug)
+- Batch concurrency limit exceeded — parallel markers have a limit of 50 items. For complex objects, use fewer than 50 due to a 2-second timeout risk
+- NoneType errors in parallel steps — bug where `file.mime_type` is None causes crashes when `.lower()` is called (fixed in later releases)
+- Input collection variable is not of type array
+
+What to look for:
+- Check the expression language (JS vs C#) for the marker input
+- Check the size and complexity of the input collection
+- Check for null/None values in collection items
+
+## Investigation
+
+1. Check if the marker input collection uses JS or C# expressions
+2. Check the collection size (>50 items for complex objects may timeout)
+3. Check for null values in collection item properties (e.g., `file.mime_type`)
+4. Verify the input collection variable is of type array
+
+## Resolution
+
+- **If JS expression error:** switch to C# expressions (see [marker-invalid-cast](marker-invalid-cast.md))
+- **If collection too large:** reduce batch size below 50 for complex objects; simple items (strings, file references) support larger lists
+- **If NoneType error:** add null checks for properties like `mime_type` before calling methods on them
+- **To set up a loop:** add a multi-instance marker on the task node, configure the input collection (must be array type), use the iterator variable on inputs within the task

package/skills/uipath-troubleshoot/references/products/maestro/playbooks/no-message-events.md

@@ -0,0 +1,51 @@
+---
+confidence: high
+---
+
+# No Message Events Found (400)
+
+## Context
+
+What this looks like:
+- HTTP 400 at Message Start Event or Message Receive Event
+- Error message: `No Message events found. Generate a trigger event that matches the expected format and try again.`
+- Variants name a specific source — `No File events found` for OneDrive/SharePoint file triggers
+
+What can cause it:
+- The trigger event was never generated at the source (e.g., no new file in SharePoint)
+- The event was generated but does not match the configured connector/action/object filter
+- Integration Service connection backing the trigger is misconfigured or expired
+- Debug mode does not always exercise triggers — known design-time limitation for SharePoint and similar triggers
+
+What to look for:
+- Message Start Event's Implementation: connector, action ("Wait for connector event"), event type
+- Whether the user is debugging vs running the deployed workflow
+- Whether the inbound event matches the configured filter (folder, file extension, mailbox, etc.)
+
+## Investigation
+
+> Substitute `<type>` with `bpmn`, `flow`, or `case` per the [Maestro investigation guide](../investigation_guide.md) § Determine the Maestro process type.
+
+
+1. Get the incident: `uip maestro <type> instance incidents <instance-id> -f <folder-key> --output json`
+2. Inspect the BPMN to confirm the Message Start Event configuration: `uip maestro <type> instance asset <instance-id> -f <folder-key> --output json`
+3. Check the Integration Service connection's health in Orchestrator UI → **Integration Service > Connections**
+4. Trigger a test event from the external system that meets the configured filter; confirm via the connector's event log
+5. If user is in debug, ask whether the same trigger works in the published flow
+
+## Resolution
+
+- **If trigger config wrong:** set the Implementation to "Wait for connector event"; pick the right connector and event type
+- **If connection expired:** re-authenticate the IS connection
+- **If filter mismatch:** generate an event that satisfies the filter (correct folder, file type, channel)
+- **If debug-only limitation:** publish the workflow and run it deployed; design-time debug has known limitations for some triggers
+- **If using a custom webhook:** regenerate the webhook URL on the connector and re-register it in the external system
+
+## Notes
+
+- This error is **Not Troubleshootable** from PIMS API alone — incident lacks the trigger configuration details and the inbound event format. Expect to ask the user for the failing element name and to share the trigger config
+
+## References
+
+- [Docs: Events in BPMN](https://docs.uipath.com/maestro/automation-cloud/latest/user-guide/events-in-bpmn-modeling-perspective)
+- [Docs: Events](https://docs.uipath.com/maestro/automation-cloud/latest/user-guide/events)