DID-ACTO / FIELD GUIDE

Troubleshooting and expectations

Use the symptom to choose a useful next step.

I cannot sign in or reset my password

Confirm you are using your dedicated DiD-Acto credentials. Complete a required first-login password change and, for administrators, authenticator enrollment. If you forgot the password, the account is disabled or enrollment cannot be completed, contact an administrator.

An administrator can reset the existing account’s password; password-reset emails are not currently available.

The portal opens but the assistant cannot connect

Check the exact https://prolug.asuscomm.com/mcp URL, the configured client ID, and that the connector is enabled for your conversation. Complete the client’s separate OAuth sign-in. An organization may require an owner to enable a connector.

Claude’s remote connector needs the service to be reachable from Anthropic’s infrastructure, not only from your computer. A working portal session does not prove the client has a valid MCP session.

My upload failed or stopped

Confirm the format and 200 MB limit, and check remaining workspace space. For an interrupted transfer, reselect the same unchanged file and use Upload or resume. A changed file is a different upload.

If the file uploads but inspection fails, examine the table structure, column types and expanded size. The parser also checks the expanded size and table structure; use the error details to identify the needed change.

The assistant cannot find my file

Copy the actual upload identifier or server path from the portal. A workstation path is not available on the server. Check that the portal and MCP client are signed in to the same account. Administrators should also check their selected workspace.

Planning needs more information

Read the specific missing assumption, variable role or support issue. Answer it from your study knowledge, or acknowledge that it is unknown. A refusal can indicate that the requested design is outside the current supported profile.

Keep unknown facts explicit. Ask the assistant whether the next step is to clarify the intervention history, repair the data, narrow the target window or use a different supported design.

My analysis is queued or taking longer than expected

The service runs one heavy fit at a time. Record the run identifier and check that same run. Sample size, resampling and sensitivity requirements affect duration. Avoid repeatedly submitting new runs while the existing one is still active.

If the run remains unchanged unexpectedly, give the administrator the run identifier and approximate submission time. Include the last reported status so the administrator can distinguish a queued job from a stalled one.

My result or report is unavailable

Confirm the selected workspace and refresh the library. A run may still be incomplete, may have failed, or may be quarantined by integrity checks. Historical records can require compatibility or integrity reconciliation.

Ask an administrator to investigate the record. Include the run identifier and the error shown; the administrator can check the saved record and recovery options.

The result does not show the effect I expected

Inspect the target, scale, uncertainty and support. Consider whether the outcome, comparison and assignment story answer the original question. When the interval is wide, ask which effect sizes remain plausible and whether more observations or a better comparison would answer the question.

If a substantive error requires a new specification, document the reason and plan a new run. Keep the earlier result so the reason for the change remains clear.