How to identify neighbors with some conditions

I'm trying to set up a toy model of connected turtles. Every turtle has a random belief [0,1] and I want to implement some conditions:

If it's a turtle with belief > .5 the turtle doesn't do anything.
If it's a turtle with belief < .5 and one of their immediate neighbour has a belief > .5, their new belief (of the turtle with belief < .5) would be belief + .1.
If the turtle with belief < .5 doesn't have an immediate neighbour with belief > .5, the turtle doesn't do anything.

The basic setup:

turtles-own [belief]

to setup
  ca
  reset-ticks
  crt 5 [
   set shape "person"
   layout-circle turtles 5
   set belief random-float 1
   ]

ask turtles [ create-links-with min-n-of 2 other turtles [distance myself] ]

end

This creates a 5 turtles cycle network. Now they have to check the beliefs of their (two) closest neighbors. I tried this:

to go

ask turtles with [one-of link-neighbors with [belief > .5]] [

set belief belief + .1
]

end

So the basic idea is that every turtle with a connected neighbour with belief > .5 has to add .1 to its own belief. I have tried several variations of the ask turtles with line, but it always shows me a runtime error of this sort: WITH expected a true/false value from (turtle 0), but got (turtle 2) instead. Any ideas? Thank you in advance.

Solution

one-of returns an agent. It is not suitable in this case for evaluating a condition with with- the condition is incomplete, since the line ask turtles with [one-of link-neighbors with [belief > .5]] translates (roughly) to "ask turtles with the condition turtle x to do something".

I think what you're after is something that uses count to count the number of believing turtles or any? to evaluate if an agentset exists:

to go-1
  ask turtles with [ (count link-neighbors with [belief > 0.5]) > 0 ] [
    set belief belief + 0.1
  ]
end

to go-2
  ask turtles with [ any? link-neighbors with [ belief > 0.5 ] ] [
    set belief belief + 0.1
  ]  
end