fix: clear error and namespace handling when triggering scheduled task jobs

[fedemaleh]

Summary

Fixes the scheduled task job trigger action, which failed with an opaque error when the scope had no active deployment and had K8S_NAMESPACE handling bugs.

• Clear error on undeployed scope — triggering a job on a scope that is not deployed now fails with a "deploy the scope first" message instead of an opaque error.
• Logging via workflow step — the trigger loads the log function through a load logging workflow step ($OVERRIDES_PATH/logging) instead of assuming it exists.
• K8S_NAMESPACE under set -u — guard the default so the script does not abort when the variable is unset; resolve from the provider namespace with a fallback.
• values.yaml in trigger-job workflow — include $SERVICE_PATH/values.yaml so K8S_NAMESPACE is provided to the trigger step.

Test plan

• bats scheduled_task/scope/tests/trigger.bats → pass (namespace resolution, set -u safety, workflow includes/logging)