Agreeable people are more likely to shift their expressed views to agree with those they are talking to.

If they're more likely to shift their views, we call them "gullible", not "agreeable".

But this is a distinction you can't apply to language models, which don't have views.