Persist source ref protection on flow workloads

What does this MR do and why?

Workload pipelines run on an internal refs/workloads/* ref, so Ci::Pipeline#protected_ref? was always false for them and protected CI/CD variables could never reach a flow.

• Ci::Workloads::WorkloadBranchService reports whether the source ref is protected.
• Ci::Workloads::RunWorkloadService stores it on the workload (protected_ref, column from MR 1).
• Ci::Pipeline#protected_ref? uses the stored value for workload-sourced pipelines.
• Ai::DuoWorkflows::StartWorkflowService passes it through.

No extra variables are exposed yet: StartWorkflowService requests no CI/CD variables, and flow triggers keep the default false.

Database: one extra single-row query for workload-sourced pipelines, memoized, on p_ci_workloads_pipeline_id_idx (pipeline_id, partition_id):

SELECT "p_ci_workloads".* FROM "p_ci_workloads"
WHERE "p_ci_workloads"."pipeline_id" = 2809347353 AND "p_ci_workloads"."partition_id" = 105 LIMIT 1;

The existing workload INSERT gains the protected_ref boolean column.

References

• Work item: #602887
• Split of !252729 (the original, now the final MR):

#	Merge request	Base
1	!254470 (merged) Add columns (merged)	master
2	!254471 (merged) Parse variables in agent-config.yml (merged)	master
3	!254472 Persist source ref protection	master
4a	!254473 Expose allowed_in_workloads on APIs	1
4b	!254474 Settings checkbox	4a
5	!252729 Gate workload variables (feature becomes active)	3, needs 2

Screenshots or screen recordings

Not applicable: no UI.

How to set up and validate locally

1. In a project, protect a branch release (Settings > Repository > Protected branches).
2. Start a flow from release (for example from the Duo Agentic Chat with that branch checked out, or with source_branch: release on the start API).
3. In a Rails console, Ci::Workloads::Workload.last has protected_ref: true and Ci::Workloads::Workload.last.pipeline.protected_ref? is true.
4. Start a flow from an unprotected branch: both are false.
5. A regular (non-flow) pipeline on release still reports protected_ref? as true.

MR acceptance checklist

Evaluate this MR against the MR acceptance checklist. It helps you analyze changes to reduce risks in quality, performance, reliability, security, and maintainability.