Separate silent issue metadata sync from public retest guidance so maintainer edits and reporter-facing comments do not share the same workflow boundary. Extract the shared parsing and classification logic into a local helper with a small node test so both workflows stay in sync without duplicating the script.
