I would agree that 2x/wk is insufficient OG, especially where the two days are consecutive. Ideally, OG should be 5x30 per week, or even 5x60 per week (though sometimes kids get burned out on the longer sessions). There is research that suggests 2x30 can be effective, but that is probably more applicable to mild/moderate dyslexia. If 5x60 is not an option, I would prioritize more frequent sessions over lengthier ones.
Yes, the teacher should have some level of access to lists of nonsense words. Here's a link to a program that generates them in a variety of languages, including English:
http://crr.ugent.be/programs-data/wuggyIt's also on this page that lists a number of pseudo word resources:
http://www.lcampanelli.org/nonword_databases/On the goal/objective: consider specifying which phonological skills he should be mastering, instead of the percentage of accuracy. E.g., "will demonstrate mastery of the three most common spellings of long a in 3 out of 4 opportunities", "will correctly spell consonant-le patterns at 90% accuracy". The OG teacher can also include a number of specific patterns (whatever she thinks is achievable in the next year). They don't have to be comprehensive of what she will work on with him, just indicators of his progress.
Sounds like he is still struggling with language pragmatics, but not pure language. You may be correct that his history of language development delays was more a function of oral-motor skills than of cognitive language. I wonder if his variable performance on social language reflects the same kind of inconsistency of application that appears to be in play for spelling--which again suggests that his success in the area relies on conscious cognitive processes, rather than automatized ones.