SPYGLASS is good.
Really needed depending on the complexity of design. You can catch Gate level simulation issues at RTL stage it self. It will reduce ECO cycle.
If you have a single clock domain you may not need. but once you have multiple clocks then for Clock Domain Crossing run is Required
How should requirements / constraints be developed for SpyGlass? What language? WHo should define the rules of checks? Where should they be taken from?
- - - Updated - - -
How do Cross Domain checks are performed? What's the design checked for?
Who should define DFT rules? What exactly does SpyGlass check in terms of DFT?