A c=0 sample plan is one in which you do a certain number of samples

. If you have zero failures (c) in those n samples, the test passes. If you have one or more failures, the plan fails. Basically, you stop sampling on the first failure.
Some "famous" c=0 plans include a sample size of 29 (if there are zero failures, I am 90% sure that the failure rate is less than 10%) and 59 (if there are zero failures, I am 95% sure that the failure rate is less than 5%).
Other plans allow you to have a couple of failures out of the number of trials. This reduces false alarm rates but requires a bigger sample size for the same AQL. A c=2 n=100 plan says I sample 100 items, and pass on 0, 1 or 2 failures, reject on 3 or more failures.