Brings existing SQL Server Agent jobs (and optionally their run history) into Minion Agent instead of rebuilding them by hand -- either by connecting live to a server's msdb, or by pasting a script you generated yourself in SSMS. Either way, the job list parses automatically as soon as the script settles -- there's no separate Parse button to click.

Target Server is the destination every parsed job gets imported to, on Commit -- it's not necessarily where the script came from. When you Live Connect (below), this app also happens to use it as the server to scan, since importing jobs onto a different box than the one you scanned them from would rarely make sense.

On Duplicate Job Names decides what happens automatically when a parsed job's name already exists in Minion Agent, so you don't have to click through each collision one at a time: Skip leaves the existing job exactly as it is and the parsed one just stays unresolved (never imported); Replace trash-cans (soft-deletes) the existing job so the name is free, then imports the parsed one in its place; Rename leaves the existing job active but renames it out of the way (appending a timestamp) so the parsed job can take its name, letting you keep both to compare. This applies the moment jobs are found, before you've looked at any of them -- a job card's own Replace/Rename buttons still let you override it for one specific job afterward.

Connect As picks who the live scan connects as to read the target server's own msdb -- either (Worker's own identity) (Worker's Windows service account) or one of your stored credentials. This is a read-only inventory action, not a job-execution credential -- it never gets attached to anything you import; every step still needs its own credential chosen separately below.

Import Run History (None / Last 30 Days / Last 90 Days / Last 6 Months / Last Year / All Available) only applies to a live scan -- pasting a script has no history to bring along. Whatever range you pick comes back with the scan and rides along to Commit, landing in the same run-history tables and reports every normal run already writes to, keyed by job/step name.

Scan for Jobs connects, reads every job in that server's msdb (plus history if requested), and parses the result automatically the moment it comes back -- no separate Parse click needed.

Paste a script instead... is the other way in: click it to reveal a plain text box, then paste whatever SSMS's "Script Job as -> CREATE To" produced (one job or several at once). Typing/pasting triggers an automatic parse about a second after it settles, so there's nothing to click once you stop typing.

Once a scan or paste actually finds jobs, the setup fields above collapse behind a one-line summary (target server + job count) so the checklist and detail area below get the room instead -- click the arrow next to that summary any time to expand or collapse it again, freely, in either direction.

Every discovered job name shows up, checked, in the checklist on the left -- uncheck any you don't want, or use the filter box to find specific names first (the All/None buttons apply to whatever the filter currently shows, not the whole list). Nothing unchecked ever gets imported, no matter how many other jobs on the same scan are otherwise resolved and ready.

Full review detail only ever appears on the right for jobs actually checked in the list -- picking through a long scan doesn't mean scrolling past detail for jobs you don't want. Each job's card shows every step (with its subsystem, and a credential dropdown that's never pre-selected, even when a name match was found -- every step needs its own deliberate pick) and every schedule, plus, for a still-unresolved duplicate, its own Replace Existing Job / Rename Existing, Then Import buttons (the same two choices On Duplicate Job Names makes automatically, offered here per-job).

The quick-assign a credential to every unresolved step row (only shown once something's been parsed) is a shortcut for large imports -- it sets every step, in every job, to run as one chosen credential in a single click, after an explicit warning. It's not how you should normally do this: least-privilege means each step gets its own credential, individually reviewed. Use it to get a big import moving, then go back and fix the steps that matter.

Import Resolved Jobs only imports jobs that are both checked and fully resolved: not a pending duplicate, every step has a real assigned credential, and every schedule has a resolvable day/time (or is a startup schedule). Anything left unresolved or unchecked simply stays in the list for next time -- nothing partial gets imported. If Commit finds a name collision that appeared after Parse (someone else created a same-named job in the meantime), that job comes back flagged as a duplicate instead of failing the whole batch, with the same Replace/Rename choice available on its card.

Related: Credentials -- what a step's credential dropdown is actually choosing between. Job Scheduling -- how an imported schedule maps to a recurring run. Gates -- not something SQL Agent has, so nothing here imports one.

See also: Glossary for related terms.