For HIV, the risk per CIM with someone who is HIV+ (the large majority of clients aren't going to be*) is somewhere between 2 in 10,000 and nil. The higher figure came from a US study of gay men in three cities that looked at what they did over a period of time and who got infected. Because oral is the (almost) universal sex act between men, they had to do some assumptions that not everyone agrees with in order to come up with the final result, and it is definitely at the high end of possible reality - there are plenty of studies saying the risk is much less and quite possibly nil. Without CIM, any risk is so near nil that anyone unlucky enough to catch it that way would probably be struck by lightning then crushed by a steamroller on the way to be tested.**
For HepB, you want to be vaccinated - it's more infectious than HIV.
For things like gonorrhoea and chlamydia, it's throat swab time - the bacterials are more infectious still and outside the urethra most cases don't have any symptoms.
* Although the infection rates for men who have sex with men are significantly higher than the general population, this is still true for people without vaginas: 'about' one in eight gay men in London are HIV+ = seven out of eight are HIV-.
** Ditto the risk for giving oral to an HIV+ woman.