Skip to main content
Start by identifying whether the failure is local input, authentication, service connectivity, or a hosted run.

​
Installation or command not found

If the command is unavailable, reinstall LoomLoom and ensure its install directory is on PATH. Confirm the installed version before comparing its command set with these docs.

​
Authentication, server, or network error

Check that:
  • LOOMLOOM_SERVER uses HTTPS and points to the intended service;
  • LOOMLOOM_TOKEN is set and belongs to that same service;
  • your network can reach the configured host;
  • the installed CLI and service are compatible.
Never paste the token into a support request.

​
Local file or schema error

Read the error first. Correct a flag, JSON/JSONL file, workbook, or TemplateSpec locally and retry the relevant validation command. doctor is not necessary for an ordinary local input error. For official workbooks:
For private templates:
If TS-TOPOLOGY-001 appears, remove new bindMode=expanded authoring. Use one workbook row per independently processed item or define a fixed number of parallel steps with dependsOn and upstreamBindings.

​
Template or model is unavailable

Availability depends on the current environment and permission. Use IDs returned by the service; do not guess them from display names or a static list.

​
Workbook validation fails after a template change

Download a fresh workbook for the current official template or exact private template version. Do not assume an older workbook remains compatible.

​
A submission timed out or returned an ambiguous error

Do not immediately repeat a paid or state-changing command.
For Market calls, also inspect the usage record. For publishing changes, inspect the Listing and review request. Reuse a client request ID only if the original payload and confirmation are unchanged.

​
Where to report a reproducible issue

Open an issue in the LoomLoom GitHub repository. Include the CLI version, operating system, sanitized doctor output, the command shape, and the error message. Remove tokens, Customer Content, signed URLs, and other sensitive values.