With Liljegren gone, we can probably accrue enough cap space to add Granlund at the deadline, even at full cap hit.
The Leafs are still running LTIR with Jarnkrok out.
Accruing cap space will help a lot. Even with Jarnkrok off of LTIR they should be able to be decently below the cap. With that said, the Leafs could get a 3rd team to retain on Granlund for a pick if needed. He's a pending UFA so typically it's not overly hard/expensive to get done.
As for Frost, my only concern is that he's another perimeter pass first player. He'd add speed and he's talented, so maybe the Leafs strike gold if the price is right. A small bonus is that he was a big Maple Leafs fan growing up.
The Leafs 23 man roster as it sits today is a shade under $85m. They currently have Hakanpaa ($1.47m), Jarnkrok ($2.1m), Dewar ($1.18m), and Mermis ($.775m) in addition to that. The total of those 4 guys is $5.525m.
In order to accrue cap space, you need to be under the cap when adding up all players on your roster + LTIR. If you assume Myers ($.775m), Benning ($1.25m), Holmberg ($.8m), and Mermis ($.775m) down to the AHL or out, you ge down to about $87m total.
HOWEVER, in order to get to that number, you actually need to remove those guys from the roster. They're not going to demote Holmberg until Jarnkrok is back. They're not allowed to demote Mermis until he's cleared to play. I don't believe Jarnkrok is even practicing. It's probably a fairly safe assumption to think that somebody's going to go on LTIR before Jarnkrok comes back, meaning they won't need to send down that 4th guy and risk waivers, meaning they probably won't accumulate much if any cap space.
Granlund's also going to be a very hot commodity come deadline time.
As for Frost, call me crazy, but I don't believe this is the Leafs year, they need to think a little beyond this year, and Frost would be a really solid long term addition. The challenge of course, while he's had a really bad start to the year, he's also Philly's 1st line centre and only centre with real offensive upside.