A cluster of topics around rogue AI agents, a Check Point exploit, slopsquatting, and ClickFix lures points to one hard truth: attackers keep aiming at trust boundaries, not just code flaws.