> ## Documentation Index
> Fetch the complete documentation index at: https://deepline.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Troubleshooting

> Ask your agent to identify the failed boundary without blind retries.

[Agent guide: CLI and SDK instructions →](/docs/troubleshooting)

## 1. Have your context ready

Keep the expected result, observed error, working folder, and any saved run identity. Remove credentials from anything you share.

## 2. Ask your agent

Paste this into your agent, then provide your task and example inputs when it asks:

```deepline-prompt theme={null}
/deepline-gtm Diagnose this Deepline failure in the environment where it happened. Separate installation, workspace authorization, provider access, invalid input, execution, usage, and destination errors. Inspect existing work before retrying. Explain which check failed, what evidence supports that conclusion, and the smallest next action. Do not interpret a timeout as no work, missing usage as zero, or an access failure as an empty result.
```

Complete any requested sign-in yourself. Give the agent the missing information before approving execution. Provider-backed work can incur usage.

## 3. Review the result

The report names the failed boundary and preserves uncertainty where evidence is missing. It gives a recovery action rather than a new execution with unexplained duplicate risk.

Ask your agent:

```deepline-prompt theme={null}
/deepline-gtm Compare the result with the acceptance rule we agreed. Show the evidence for accepted and rejected outcomes, unresolved questions, and recorded Deepline usage where work ran. If work is still running or a response is uncertain, inspect the original saved execution before proposing another request. Explain the next decision I need to make.
```

Approve a larger job, recurring work, or external changes only after reviewing that decision.

## Check your understanding

A provider request timed out. Can you safely report no result and submit a new execution immediately?

> **Show the answer**
>
> No. Preserve the original execution identity and investigate its outcome. The
> request may have completed even though the response was lost.

## Related guides

* [Multi-org setup](/docs/human/sdk-v2/multi-org)
* [Validate Incoming CSV Files Before CRM Enrichment](/docs/human/guides/validate-incoming-csv-files)
* [Run logs and events](/docs/sdk-v2/run-events)
* [Recover a tool execution](/docs/human/sdk-v2/tool-execution-recovery)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.