Add more Nexus operation timeouts.

[Quinn-With-Two-Ns]

Add more Nexus operation timeouts.

---

Note

Medium Risk
Touches Workflow/worker Nexus command plumbing and failure decoding, which can affect how production workflows observe operation failures; changes are bounded and covered by new integration tests.

Overview
Adds Nexus operation scheduleToStartTimeout and startToCloseTimeout options to the Workflow Nexus client/interceptor API and wires them through to the scheduleNexusOperation command.

Fixes decoding of Nexus operation timeouts in WorkflowCodecRunner by reading the timedOut failure field (was incorrectly using cancelled), and adds integration tests that assert SCHEDULE_TO_START and START_TO_CLOSE timeouts surface as NexusOperationFailure causes.

CI is pinned to a newer TESTS_CLI_VERSION to support these timeout behaviors.

^{Written by Cursor Bugbot for commit 9eb0f94. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

[cursor]

Pinned CLI version may be accidentally committed

Medium Severity

TESTS_CLI_VERSION was previously commented out (the normal state, matching release.yml), but is now uncommented and pinned to a specific version. The comment above says "Use these variables to force specific version," implying this is an override mechanism, not a permanent setting. This pin affects all CI tests globally, not just the new Nexus timeout tests. If left in, it prevents CI from testing against newer CLI/server releases and may mask compatibility issues.