Improve network error detail and timeout behavior

[dubscode]

Summary

Please improve error reporting and timeout/process-exit behavior for failed Linear API requests.

Observed friction

During a real workflow, this command returned a very generic fetch error:

linear issues list --team BB --json --limit 100

Output:

{
  "ok": false,
  "entity": "issues",
  "action": "list",
  "error": {
    "code": "UNKNOWN",
    "message": "Fetch failed",
    "details": {}
  }
}

One list process also appeared to keep running after the failure, requiring process inspection/cleanup.

Requested improvements

• Include underlying network error details when safe: HTTP status, response body, DNS/socket timeout reason, or SDK error code.
• Add a bounded request timeout.
• Ensure failed commands always exit promptly.
• Include retry guidance only when the error is likely transient.
• Keep JSON errors machine-readable and stable.

Why this matters

Generic Fetch failed output is hard for both humans and agents to act on. It does not distinguish auth problems, workspace/team resolution problems, DNS/network issues, Linear API failures, or local runtime errors.

Better error detail would reduce unnecessary retries and make bug reports much easier to write.