But I agree with you, there are incentives to not share the best prompt injection attacks.
Even in LLM jailbreak CTFs I've seen, it ends up feeling like underpaid work when it's sponsored by Microsoft and the prize pool is, say $10k (including stuff like azure credits) considering the salaries AI safety engineers command at big tech!