SP (4), in order:
RP (7), in loose order of usage in high-leverage situations:
Starting Lineup (9), in batting order:
RF Betts R
LF Benintendi L
DH Martinez R
SS Bogaerts R
1B Moreland/Pearce (I'd go Pearce unless Moreland shows signs of life)
2B Kinsler R
3B Nunez R
C Leon S
CF Bradley Jr. L
1B/OF Pearce (starts against lefties for sure, and maybe even righties)
I would not be opposed to moving Bradley Jr. higher in the lineup as well, given how well he's hit the last two months (and really, he's been making good contact since before then). If they go with Pearce at 1B, they could use Bradley to break apart the string of righties from 3rd through 7th in the above lineup. The catcher hitting right before Betts is not ideal, but hopefully they would pinch-hit frequently at that spot if they have the bench above.
Rationale regarding the last two pitchers and last two hitters:
Workman and Poyner have earned the spots of late (and have solid numbers overall). They don't inspire great confidence, but more so than other options. They hopefully will not face high-leverage situations, except maybe Poyner in 1-batter LOOGY outings.
By keeping three catchers, it gives the team the option of pinch-hitting for Leon early in a game (e.g. even 3rd or 4th inning) if his turn comes up in a high-leverage scenario. By having multiple decent PH options (Devers, Holt, one of Pearce/Moreland, and Swihart), they could pinch-hit twice in a game for the C if need be.
Devers provides insurance against Nunez's knee while also giving them a pinch hitter with decent power. I don't want to see him on the field much, though.
1) Tzu-Wei Lin: Would provide late-inning defensive help at 3B, but Nunez has been average-ish/passable there. That doesn't seem a big enough reason for a roster spot when Lin can't be counted on as a hitter and is not necessarily a better pinch-runner than Holt or Swihart. He can back up 2B/SS/CF as well, but the plan would be for the starters there to play every inning (and other backups at those positions are already on the roster). Having a PH with some power like Devers seems more likely to make a difference.
2) Hembree: The whole idea that somehow Hembree was great with inherited runners was probably just a small sample bias; there's nothing in his repertoire or past history to suggest he should be all that great in those situations, and regression to the mean has set in heavily.
3) Kelly: Kelly's velocity is great but we've seen enough to know it's straight and hittable all too frequently. I believe some have suggested that when he's on a hot streak, he mixes in other pitches more (or at least in a better way); if that's the case, he certainly isn't doing so right now.
4) Velazquez: The 4.32 FIP, 1.47 WHIP and low K-totals are more descriptive of his performance than his W-L or ERA. His swing/long role will be occupied by better pitchers in the playoffs (Wright/Eovaldi).
5) Johnson: Similar story as Velazquez above. Poyner's splits make him a slightly better LOOGY option than Johnson.
6) Pomeranz: Ugh.